Archive-Date: Mon, 01 Nov 1993 14:14:11 CST Sender: list-mgr@WKUVX1.BITNET Date: Mon, 01 Nov 1993 08:24:04 EDT From: George Weitzel Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00974E18.ECFD1660.21705@ube.ub.umd.edu> Subject: VMSMail Patch for Pathworks Mail? Speaking of the VMSMail Patch, has anyone patched PCFS_MAILSHR.EXE so that Pathworks Mail users can also avoid the MX% syntax? =================================================================== George H. Weitzel BITNET: george@UBE Network Services Manager Internet: george@ube.ub.umd.edu Academic Computing Center Voice: (410) 234-3942 University of Baltimore FAX: (410) 539-3714 1420 North Charles Street Baltimore, Maryland, USA 21201 =================================================================== ================================================================================ Archive-Date: Mon, 01 Nov 1993 14:49:31 CST Sender: list-mgr@WKUVX1.BITNET Date: Mon, 01 Nov 1993 15:46:15 EDT From: George Weitzel Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <00974E56.B2C64000.24094@ube.ub.umd.edu> Subject: Re: VMSMail Patch for Pathworks Mail? > Speaking of the VMSMail Patch, has anyone patched PCFS_MAILSHR.EXE so that > Pathworks Mail users can also avoid the MX% syntax? > > > =================================================================== > George H. Weitzel BITNET: george@UBE Actually, I may have listed the wrong executable. It might be PCSA_MAILSHR.EXE instead (or both?). ^^^^ =================================================================== George H. Weitzel BITNET: george@UBE Network Services Manager Internet: george@ube.ub.umd.edu Academic Computing Center Voice: (410) 234-3942 University of Baltimore FAX: (410) 539-3714 1420 North Charles Street Baltimore, Maryland, USA 21201 =================================================================== ================================================================================ Archive-Date: Tue, 02 Nov 1993 08:09:49 CST Sender: list-mgr@WKUVX1.BITNET From: "Pasztor Miklos" Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Date: Tue, 2 Nov 1993 14:00:42 CET Subject: Automatic loop detection Message-ID: <534D419C9@aszi.sztaki.hu> MX 3.1 release notes states: # 3. A Received: line counter has been added to the # SMTP, DECnet-SMTP, and X.25-SMTP servers, in order # to detect mail loops. When MX sees 10 Received # headers from the local system in a message, it # assumes that the message is looping and does not # accept the message. We run MX 3.3, and due to some mailer misconfigurations (certainly always at someone other's site :-) sometimes we have to hunt for messages which are in loop. I defined a route intentionally to cause a loop (actually DEF PATH LOOP DECNET/ROUTE=), sent a message to xx@loop, and the message looped forever. Am I missing something ? Thanks, Miklos ==================================================================== Pa'sztor Miklo's | E-mail: pasztor@hugbox.bitnet MTA SZTAKI/ASZI Budapest Victor H. u. 18-22 | Phone: (36)-(1)-149-75-32 Institute for Computation and Automation, Hungarian Academy of Sciences ==================================================================== ================================================================================ Archive-Date: Tue, 02 Nov 1993 14:48:55 CST Sender: list-mgr@WKUVX1.BITNET Date: Tue, 02 Nov 1993 15:36:45 -0500 (EST) From: "Jonathan E. Hardis" Reply-To: MX-List@WKUVX1.BITNET Subject: Cancellation Bug To: MX-List@WKUVX1.BITNET CC: hardis@garnet.nist.gov Message-ID: <00974F1E.8993FF60.26982@garnet.nist.gov> Content-Transfer-Encoding: 7BIT I just noticed an interesting effect. Someone sent a "subscribe" message to my mailing list, instead of to the "-request" address. I received it fairly quickly, since I got it on LOCAL delivery. So, as a courtesy to my SMTP subscribers, I issued an MCP QUEUE CANCEL on that message -- no reason that they need see it. What happened was that the SMTP agent continued to try to deliver the message to everyone! However, I got a large log file back that said %RMS-F-FNF, file not found each time it tried to send the message to someone else. I don't think any harm was done, but should there be a more graceful way of aborting a message being sent to a mailing list? - Jonathan ================================================================================ Archive-Date: Tue, 02 Nov 1993 15:18:38 CST Sender: list-mgr@WKUVX1.BITNET Date: Tue, 02 Nov 1993 16:04:11 EST From: "Brian Tillman" Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <00974F22.5E5F5F20.8310@swdev.si.com> Subject: Analog to MULTIPLE_FROM The current version of MX allows a person to reply to a message that specified multiple addresses in the From: or Reply-To: headers. While this may be handy in some cases, I feel there is another feature that would be handier. That is, if a person sends a message to me and others, I'd like to be able to respond to everyone the original person addressed. That is, if the other person (suppose his name is FRANK) sends to MX%TOM, MX%DICK, MX%HARRY, and me, I'd like to REPLY to that message and have the VMS From appear as MX%FRANK, MX%TOM, MX%DICK, and MX%HARRY. Alternatively, it would be acceptable if the reply went to FRANK and the others to whom FRANK wrote would appear in the CC list of my reply to FRANK. Ideally, I'd like MX to handle CC addresses as well, so that those people who were CC'd on FRANK's mail to me would appear in the CC list of my reply to FRANK, but that is of secondary utility to me. If this is determined to be feasible, please add it to the wish list. Thanks. -----------------------------+-------------------------------- Brian Tillman | Internet: tillman@swdev.si.com Smiths Industries, Inc. | tillman_brian@si.com 4141 Eastern Ave., MS129 | Hey, I said this stuff myself. Grand Rapids, MI 49518-8727 | My company has no part in it. -----------------------------+-------------------------------- ================================================================================ Archive-Date: Tue, 02 Nov 1993 15:38:58 CST Sender: list-mgr@WKUVX1.BITNET Date: Tue, 02 Nov 1993 15:36:51 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00974F1E.8CBAD1D2.14906@WKUVX1.BITNET> Subject: RE: Cancellation Bug "Jonathan E. Hardis" writes: > >I just noticed an interesting effect. > >Someone sent a "subscribe" message to my mailing list, instead of to the >"-request" address. I received it fairly quickly, since I got it on LOCAL >delivery. So, as a courtesy to my SMTP subscribers, I issued an MCP QUEUE >CANCEL on that message -- no reason that they need see it. > >What happened was that the SMTP agent continued to try to deliver the >message to everyone! However, I got a large log file back that said > %RMS-F-FNF, file not found >each time it tried to send the message to someone else. > >I don't think any harm was done, but should there be a more graceful way of >aborting a message being sent to a mailing list? > Probably. ;-) Added to the ever-growing wish list.... Thanks! Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Tue, 02 Nov 1993 15:41:05 CST Sender: list-mgr@WKUVX1.BITNET Date: Tue, 02 Nov 1993 15:39:18 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00974F1E.E4901EF3.14908@WKUVX1.BITNET> Subject: RE: Analog to MULTIPLE_FROM "Brian Tillman" writes: > >The current version of MX allows a person to reply to a message that specified >multiple addresses in the From: or Reply-To: headers. While this may be handy >in some cases, I feel there is another feature that would be handier. That is, [...] >Ideally, I'd like MX to handle CC addresses as well, so that those people who >were CC'd on FRANK's mail to me would appear in the CC list of my reply to >FRANK, but that is of secondary utility to me. > >If this is determined to be feasible, please add it to the wish list. Thanks. I'll add it to the wish list, but the limiting factor here is VMS Mail, not MX. Off the top of head, I know of no way for MX to get all that info from VMS Mail. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Wed, 03 Nov 1993 02:48:14 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 03 Nov 1993 09:44:45 +0100 From: Herve GILIBERT CRI universite Jean Monnet ST ETIENNE Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <00974FB6.875A6700.16839@stroph.univ-st-etienne.fr> Subject: About MX accounting. Hi Neters ! Can somebody explain me how MX accounting files (MX_*_ACC.DAT) are managed ? How many times they records informations ? When are they purged ? Thanks in advance. -- Herve GILIBERT UUUU UUUU CRITeR UUUU UUUU Universite Jean Monnet St-Etienne UUUU UUUU 23, Rue du Dr. P. Michelon UUUU UUUU 42023 ST ETIENNE Cedex 2 UUUU UUUU Tel : 33 77 42 15 79 UUUU UUUU Fax : 33 77 42 15 75 UUUUUUUUUUU RFC822: gilibert@criter.univ-st-etienne.fr UUUUUUUUU ================================================================================ Archive-Date: Wed, 03 Nov 1993 03:03:29 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 03 Nov 1993 03:56:20 CST From: Kenny Kon Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <00974F85.DB221E40.284@bible.acu.edu> Subject: Transferring ascii file to subscription list Date sent: 3-NOV-1993 03:55:14 Can anyone tell me how I can move a list of user names in an ASCII file ( the usernames are in the From: header format) to the *.mailing_list file? Thank you. --------------------------------------------------------------------------- Kenny Kon Internet: kon@bible.acu.edu ACU Station, Box 7725 kkon@acuvax.acu.edu Abilene, TX 79699 BITNET: kkon@acuvax.bitnet (915)672-4915 CompuServe: 72724,3547@compuserve.com (915)674-3708 (work) ================================================================================ Archive-Date: Wed, 03 Nov 1993 09:00:32 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 03 Nov 1993 08:58:42 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00974FB0.18A87050.15218@WKUVX1.BITNET> Subject: RE: Transferring ascii file to subscription list Kenny Kon writes: > >Date sent: 3-NOV-1993 03:55:14 > >Can anyone tell me how I can move a list of user names in an ASCII file ( >the usernames are in the From: header format) to the *.mailing_list file? >Thank you. > Just add ADD commands before each address and mail the file to the -request address for the list. Or, if the file contains only addresses, one per line, you can copy it to MX_MLIST_DIR:LISTNAME.MLIST (where LISTNAME is the name of the list) and MX will be able to read it---the next time a user subscribes, the file will be automatically converted to a listname.MAILING_LIST file. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Wed, 03 Nov 1993 09:13:24 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 3 Nov 1993 10:15:44 -0600 To: mx-list%wkuvx1.bitnet@ulkyvm.louisville.edu From: reef@xenon.che.ilstu.edu Reply-To: MX-List@WKUVX1.BITNET Subject: Re: Transferring ascii file to subscription list The simplest way is just to get the names from the list from the requester and then save the letter as an ASCII file. Seems straight-forward to me. That's what I do. Best, Reef *************************************************************************** Philip D. (Reef) Morse II Internet address: reef@xenon.che.ilstu.edu Department of Chemistry Telephone (voice and voice mail): (309) 438-5595 Illinois State University Telephone (FAX): (309) 438-5538 Normal, IL 61761 Telephone (answering machine): (309) 829-9257 USA *************************************************************************** ================================================================================ Archive-Date: Wed, 03 Nov 1993 14:03:52 CST Sender: list-mgr@WKUVX1.BITNET Subject: Obtaing MX Software Message-ID: <1993Nov3.142219.1@cortex> From: Reply-To: MX-List@WKUVX1.BITNET Date: 3 Nov 93 14:22:19 -0500 To: MX-List@WKUVX1.BITNET Hi! I am considiering using MX as a mail router for a Pathworks network to allow access to the internet. Does anyone know where I can obtain the software? Any help would be greatly appreciated. Thanks Peter Barrera BARRERPD@HQ.CORTEX.PROSPECT.COM FAX (617) 622 1900 PH (617) 622 1934 Cortex Corporation 100 Fifth Ave Waltham, MA 02154 ================================================================================ Archive-Date: Wed, 03 Nov 1993 14:45:08 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: Obtaing MX Software Message-ID: <1993Nov3.130349.258@buckie.hsc.colorado.edu> From: Date: 3 Nov 93 13:03:49 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <1993Nov3.142219.1@cortex>, pbarrera@cortex.prospect.com (Mike Boucher) writes: >I am considiering using MX as a mail router for a Pathworks network to >allow access to the internet. Does anyone know where I can obtain the >software? Any help would be greatly appreciated. Anonymous FTP to ftp.spc.edu in the [.MX] directory, or from fileserv@wkuvx1.bitnet (in about 150 parts) if you can't use FTP. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Wed, 03 Nov 1993 14:45:26 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 03 Nov 1993 14:31:54 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET CC: pbarrera@cortex.prospect.com Message-ID: <00974FDE.A4483F0E.15327@WKUVX1.BITNET> Subject: RE: Obtaing MX Software writes: > >Hi! > >I am considiering using MX as a mail router for a Pathworks network to >allow access to the internet. Does anyone know where I can obtain the >software? Any help would be greatly appreciated. > I'm posting this in case anyone else reading the newsgroup needs the info. The current version of MX is V3.3. It's available via anonymous ftp from ftp.spc.edu in [.MX.MX033]. The file MX033.ZIP contains the VMSINSTAL kit for MX. The sources are in MX033_SRC.ZIP. If you need UNZIP, you can find .EXEs for it in [.MX]. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Wed, 03 Nov 1993 14:45:44 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 3 Nov 1993 14:33:59 -0600 (CST) From: "H.A. Kippenhan Jr." Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <931103143359.2a8015ab@fndcd.fnal.gov> Subject: RE: Obtaing MX Software Hi Peter: | From: SMTP%"MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU" 3-NOV-1993 14:14:24.84 | Subj: Obtaing MX Software | | Hi! | | I am considiering using MX as a mail router for a Pathworks network to | allow access to the internet. Does anyone know where I can obtain the | software? Any help would be greatly appreciated. | Use anonymous FTP to access ftp.spc.edu (St. Peter's College somewhere in the wilds of New Jersey). Regards - Kipp - {~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~} { H.A. Kippenhan Jr. | Internet: Kippenhan@FNDCD.FNAL.GOV } { National HEPnet Management | HEPnet/NSI DECnet: FNDCD::KIPPENHAN } { Fermi National Accelerator Lab. | BITnet: Kippenhan@FNDCD.BITNET } { P.O. Box 500 MS: FCC-3E/368 | Telephone: (708) 840-8068 } { Batavia, Illinois 60510 | FAX: (708) 840-8463 } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ================================================================================ Archive-Date: Wed, 03 Nov 1993 22:31:51 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 03 Nov 1993 23:24:02 CST From: Kenny Kon Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <00975028.FAEC1980.825@bible.acu.edu> Subject: A few questions Date sent: 3-NOV-1993 23:18:41 Can anyone help me with a few problems I have? 1) When the ADD_MESSAGE.TXT is sent to a new subscriber, it is sent via the Postmaster's username, can this be changed to the list owner's name? 2) The mailing list I have is in all lower-case. I tried unsubscribing from a remote account but couldn't since all remote user names have to be in uppercase. Is there a quick way to remedy this problem or do I have to get a list via review and REMOVE the lowercase names and ADD the uppercase equivalent? 3) I want to change the BITNET router from CUNY to RICE since I am in Abilene, Texas. Does anyone know the exact address for RICE's BITNET router? Thanks. --------------------------------------------------------------------------- Kenny Kon Internet: kon@bible.acu.edu ACU Station, Box 7725 kkon@acuvax.acu.edu Abilene, TX 79699 BITNET: kkon@acuvax.bitnet (915)672-4915 CompuServe: 72724,3547@compuserve.com (915)674-3708 (work) ================================================================================ Archive-Date: Wed, 03 Nov 1993 22:43:53 CST Sender: list-mgr@WKUVX1.BITNET Date: Wed, 03 Nov 1993 23:35:03 CST From: Kenny Kon Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <0097502A.85599C40.827@bible.acu.edu> Subject: Disabling SMTP Date sent: 3-NOV-1993 23:30:55 Is it normal to be able to use SMTP%"address" after installing MX? Now all the mail comes in via MX, but mail can still be sent via MX%"address" or SMTP%"address" Could someone tell me how to disable SMTP so that everyone will be able to use only MX%. --------------------------------------------------------------------------- Kenny Kon Internet: kon@bible.acu.edu ACU Station, Box 7725 kkon@acuvax.acu.edu Abilene, TX 79699 BITNET: kkon@acuvax.bitnet (915)672-4915 CompuServe: 72724,3547@compuserve.com (915)674-3708 (work) ================================================================================ Archive-Date: Thu, 04 Nov 1993 06:37:27 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 06:26:44 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00975064.083BA1AB.15618@WKUVX1.BITNET> Subject: RE: Disabling SMTP Kenny Kon writes: > >Is it normal to be able to use SMTP%"address" after installing MX? Now all >the mail comes in via MX, but mail can still be sent via MX%"address" or >SMTP%"address" Could someone tell me how to disable SMTP so that everyone >will be able to use only MX%. > You need to either rename the SMTP_MAILSHR.EXE file to something else, or you can: $ DEFINE/SYS/EXEC MAIL$PROTOCOL_SMTP MX_MAILSHR That'll cause any reference to SMTP% to actually invoke MX. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Thu, 04 Nov 1993 06:42:06 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 06:24:46 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00975063.C1843999.15612@WKUVX1.BITNET> Subject: RE: A few questions Kenny Kon writes: > >Can anyone help me with a few problems I have? > >1) When the ADD_MESSAGE.TXT is sent to a new subscriber, it is sent via >the Postmaster's username, can this be changed to the list owner's name? > Actually, it goes out with the MXserver address on the From: line. This is so that the user can just REPLY to the message to signoff, etc. The Postmaster is the Sender:; that currently can't be changed, though I'll add it to the wish list. >2) The mailing list I have is in all lower-case. I tried unsubscribing >from a remote account but couldn't since all remote user names have to be >in uppercase. Is there a quick way to remedy this problem or do I have to >get a list via review and REMOVE the lowercase names and ADD the uppercase >equivalent? > That's the easiest way. Just use REMOVE/NONOTIFY and ADD/NONOTIFY and the subscriber will never know the difference. Also, MX does not uppercase remote usernames, just remote host names. MX maintains the original case of remote usernames. >3) I want to change the BITNET router from CUNY to RICE since I am in >Abilene, Texas. Does anyone know the exact address for RICE's BITNET >router? > RICEVM is the node you want, I believe. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Thu, 04 Nov 1993 09:04:08 CST Sender: list-mgr@WKUVX1.BITNET Date: 4 Nov 93 08:51:00 CST From: "John Clement-Rice U. (713)527-8101x2037" Reply-To: MX-List@WKUVX1.BITNET Subject: RE: A few questions To: "MX-List" The Rice U bitnet router is: ricevm1.rice.edu John Clement Rice University ================================================================================ Archive-Date: Thu, 04 Nov 1993 09:32:18 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 09:27:14 CST From: "George D. Greenwade" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU CC: kon@bible.acu.edu, txbitnet@utdallas.edu, lippke@UTDALLAS.BITNET, dboyes@is.rice.edu Message-ID: <0097507D.3F448BE0.26568@SHSU.edu> Subject: RE: A few questions On Wed, 03 Nov 1993 23:24:02 CST, Kenny Kon posted to MX-List@WKUVX1.BITNET: >..... > 3) I want to change the BITNET router from CUNY to RICE since I am in > Abilene, Texas. Does anyone know the exact address for RICE's BITNET > router? I think you'll need another gateway site as Rice recently left (or is very soon scheduled to leave) the BITNET core. My guess is that UTDallas is the preferred gateway at the moment, but I believe that UTD is also about to leave the core, as well. Given the movement away from BITNET in Texas (I understand that we, along with more than a few other Texas sites, are also scheduled to leave BITNET in the near future), what is the preferred routing for BITNET/Internet gatewaying in Texas now? Any rough ideas as to how long can that be assumed stable? Regards, George %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% George D. Greenwade, Ph.D. Bitnet: BED_GDG@SHSU Department of Economics and Business Analysis THEnet: SHSU::BED_GDG College of Business Administration Voice: (409) 294-1266 P. O. Box 2118 FAX: (409) 294-3612 Sam Houston State University Internet: bed_gdg@SHSU.edu Huntsville, TX 77341 bed_gdg%SHSU.decnet@relay.the.net %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% BTW: Texas BITNET and networking issues are covered by TXBITNET@UTDALLAS or TXBITNET@VM.UTDALLAS.EDU (a pretty good list generally right on topic!), which is served by normal VM/CMS LISTSERV commands at that site if you are interested in subscribing, checking out archives, etc. ================================================================================ Archive-Date: Thu, 04 Nov 1993 10:00:41 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 10:00:05 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00975081.D62B1110.15759@WKUVX1.BITNET> Subject: RE: About MX accounting. Herve GILIBERT CRI universite Jean Monnet ST ETIENNE writes: > >Hi Neters ! >Can somebody explain me how MX accounting files (MX_*_ACC.DAT) are managed ? They're just text files that MX appends records to. >How many times they records informations ? For every message processed, if accounting is enabled. >When are they purged ? Whenever you delete them. I have a job that runs on the 1st of every month that resets the accounting logs, thus creating a new log for each month. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Thu, 04 Nov 1993 10:54:49 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 17:50:58 +0100 From: Herve GILIBERT CRI universite Jean Monnet ST ETIENNE Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <009750C3.9E161080.17128@stroph.univ-st-etienne.fr> Subject: RE: About MX accounting. > Whenever you delete them. I have a job that runs on the 1st of every > month that resets the accounting logs, thus creating a new log for > each month. > It's just I want to do, but something strange happens : a new version of accounting files is created every one or two days. I would like have only one version during a month, then run MXSUM (from contrib directory) , and reset the file (mcp reset local /acc). What am I doing wrong which create news version of accounting file ? Sometimes I run MXSUM during the day, Does it matters ? Here the files created (I never reseted any agent during this period). STROPH $dir /dat mx_local_dir:mx_*.dat /dat Directory MX_ROOT:[LOCAL] MX_LOCAL_ACC.DAT;16 4-NOV-1993 13:53:11.54 MX_LOCAL_ACC.DAT;15 4-NOV-1993 00:09:19.34 MX_LOCAL_ACC.DAT;14 3-NOV-1993 16:30:25.63 MX_LOCAL_ACC.DAT;13 3-NOV-1993 10:16:46.70 MX_LOCAL_ACC.DAT;12 2-NOV-1993 23:35:30.41 Sorry, Hunter, for my poor English. Feel free to ask me more informations. Thanks in advance. Total of 5 files. -- Herve GILIBERT UUUU UUUU CRITeR UUUU UUUU Universite Jean Monnet St-Etienne UUUU UUUU 23, Rue du Dr. P. Michelon UUUU UUUU 42023 ST ETIENNE Cedex 2 UUUU UUUU Tel : 33 77 42 15 79 UUUU UUUU Fax : 33 77 42 15 75 UUUUUUUUUUU RFC822: gilibert@criter.univ-st-etienne.fr UUUUUUUUU ================================================================================ Archive-Date: Thu, 04 Nov 1993 11:01:18 CST Sender: list-mgr@WKUVX1.BITNET From: Reply-To: MX-List@WKUVX1.BITNET Message-ID: <9311041654.AA08159@is.rice.edu> Subject: Re: A few questions To: Date: Thu, 4 Nov 1993 10:54:48 -0600 (CST) CC: MX-List%WKUVX1.bitnet@ulkyvm.louisville.edu, kon@bible.acu.edu, txbitnet@utdallas.edu, lippke@UTDALLAS.BITNET, dboyes@is.rice.edu MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit > On Wed, 03 Nov 1993 23:24:02 CST, Kenny Kon posted to > MX-List@WKUVX1.BITNET: > >..... > > 3) I want to change the BITNET router from CUNY to RICE since I am in > > Abilene, Texas. Does anyone know the exact address for RICE's BITNET > > router? > > I think you'll need another gateway site as Rice recently left (or is very > soon scheduled to leave) the BITNET core. My guess is that UTDallas is the > preferred gateway at the moment, but I believe that UTD is also about to > leave the core, as well. Given the movement away from BITNET in Texas (I > understand that we, along with more than a few other Texas sites, are also > scheduled to leave BITNET in the near future), what is the preferred > routing for BITNET/Internet gatewaying in Texas now? We are currently recommending that people contact Princeton or CUNY for BITNET/Internet gateway services. I wouldn't plan on there being other than minimal BITNET presence in Texas beyond next spring -- there's just no advantage to maintaining BITNET connectivity now, especially since an overwhelming majority of the current Texas BITNET traffic is already carried over IP or DECnet circuits. > Any rough ideas as to > how long can that be assumed stable? Until April/May at the very latest. Contact Michael Gettes (gettes@princeton.edu) to arrange to use the Princeton gateway. David Boyes postmaster, ricevm1.rice.edu dboyes@rice.edu ================================================================================ Archive-Date: Thu, 04 Nov 1993 15:04:08 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 16:40:49 EST From: "Andy, Systems Manager" Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET CC: udaa055@BAY.CC.KCL.AC.UK Message-ID: <009750B9.D16C0C00.3616@bay.cc.kcl.ac.uk> Subject: Compressing the MX queue Dear All, Following recent discussions on keeping the MX queue file clean, I've enclosed a short DCL procedure which will do the job. It's based on Hunter Goatley's original but I've added some improvements, such as error checking and making it fail safe. It's designed to be run overnight (via whatever mechanism you use to auto run such things - DECscheduler, CRON etc.) and will shutdown MX, compress the file and then restart MX when done using the new shortened file. It will automatically restart MX if aborted partway, if the compression fails for any reason, or if it can't get exclusive access to the queue file within a reasonable time (configurable). I'm open to suggestions for improvements. Regards, Andy Harper Kings College London $! ------------------ CUT HERE ----------------------- $ v='f$verify(f$trnlnm("SHARE_UNPACK_VERIFY"))' $! $! This archive created: $! Name : MX_SHRINK $! By : Andy Harper $! Date : 4-NOV-1993 15:20:56.83 $! Using: VMS_SHARE 8.4, (C) 1993 Andy Harper, Kings College London UK $! $! Credit is due to these people for their original ideas: $! James Gray, Michael Bednarek $! $! TO UNPACK THIS SHARE FILE, CONCATENATE ALL PARTS IN ORDER $! AND EXECUTE AS A COMMAND PROCEDURE ( @name ) $! $! THE FOLLOWING FILE(S) WILL BE CREATED AFTER UNPACKING: $! 1. MX_SHRINK.COM;9 $! 2. MX_SHRINK.DOC;2 $! $ set="set" $ set symbol/scope=(nolocal,noglobal) $ f=f$parse("SHARE_UNPACK_TEMP","SYS$SCRATCH:."+f$getjpi("","PID")) $ e="write sys$error ""%UNPACK"", " $ w="write sys$output ""%UNPACK"", " $ if .not. f$trnlnm("SHARE_UNPACK_LOG") then $ w = "!" $ if f$getsyi("CPU") .gt. 127 then $ goto start $ ve=f$getsyi("version") $ if ve-f$extract(0,1,ve) .ges. "4.4" then $ goto start $ e "-E-OLDVER, Must run at least VMS 4.4" $ v=f$verify(v) $ exit 44 $unpack: subroutine ! P1=filename, P2=checksum, P3=attributes,P4=size $ if f$parse(P1) .nes. "" then $ goto dirok $ dn=f$parse(P1,,,"DIRECTORY") $ w "-I-CREDIR, Creating directory ''dn'" $ create/dir 'dn' $ if $status then $ goto dirok $ e "-E-CREDIRFAIL, Unable to create ''dn' File skipped" $ delete 'f'* $ exit $dirok: $ x=f$search(P1) $ if x .eqs. "" then $ goto file_absent $ e "-W-EXISTS, File ''P1' exists. Skipped" $ delete 'f'* $ exit $file_absent: $ w "-I-UNPACK, Unpacking ", P5, " of ", P6, " - ", P1, " - ", P4, " Blocks" $ n=P1 $ if P3 .nes. "" then $ n=f $ if .not. f$verify() then $ define/user sys$output nl: $ EDIT/TPU/NOSEC/NODIS/COM=SYS$INPUT/NOJOURNAL 'f'/OUT='n' PROCEDURE GetHex(s,p)LOCAL x1,x2;x1:=INDEX(t,SUBSTR(s,p,1))-1;x2:=INDEX(t, SUBSTR(s,p+1,1))-1;RETURN 16*x1+x2;ENDPROCEDURE;PROCEDURE SkipPartsep LOCAL m; LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m);EXITIF INDEX(ERASE_LINE, "-+-+-+-+-+-+-+-+")=1;ENDLOOP;ENDPROCEDURE;PROCEDURE ProcessLine LOCAL c,s,l,b, n,p;c := ERASE_CHARACTER(1);s := ERASE_LINE;EDIT(s,"TRIM");IF c = "X" THEN SPLIT_LINE; ENDIF;MOVE_HORIZONTAL(-1);l := LENGTH(s);p := 1; LOOP EXITIF p > l;c := SUBSTR(s,p,1);p := p+1;CASE c FROM ' ' TO '`' ['\']: b:= GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT( SUBSTR(CURRENT_LINE, CURRENT_OFFSET-b+1,n));['&']: b:=GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT(ASCII(n)*b);['`']: COPY_TEXT(ASCII(GetHex(s,p))); p:=p+2;[INRANGE, OUTRANGE]: COPY_TEXT(c);ENDCASE;ENDLOOP;ENDPROCEDURE;PROCEDURE Decode LOCAL m; POSITION(BEGINNING_OF(b));LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m); IF INDEX(CURRENT_LINE,"+-+-+-+-+-+-+-+-")=1 THEN SkipPartSep;ELSE ProcessLine; MOVE_HORIZONTAL(1);ENDIF;ENDLOOP;ENDPROCEDURE;SET(FACILITY_NAME,"UNPACK");SET( SUCCESS,OFF);SET(INFORMATIONAL,OFF);t:="0123456789ABCDEF";f:=GET_INFO( COMMAND_LINE,"file_name");b:=CREATE_BUFFER(f,f);Decode;WRITE_FILE(b,GET_INFO( COMMAND_LINE,"output_file"));QUIT; $ if p3 .eqs. "" then $ goto dl $ open/write fdl &f $ write fdl "RECORD" $ write fdl P3 $ close fdl $ w "-I-CONVRFM, Converting record format to ", P3 $ convert/fdl=&f &f-1 &P1 $dl: delete 'f'* $ checksum 'P1' $ if checksum$checksum .nes. P2 then $ - e "-E-CHKSMFAIL, Checksum of ''P1' failed." $ exit $ endsubroutine $start: $! $ create 'f' X$! Purpose: X$!`09Automatically compress the MX QUEUE file X$! X$! Parameters: X$!`09P1`09`5Boptional`5D delay time between queue file access retries (default V: 15 seconds) X$!`09P2`09`5Boptional`5D number of retries (default: 240 ) X$!`09P3`09`5Boptional`5D username to run MX (default: SYSTEM) X$! X$! Privileges: X$!`09 X$! Environment: X$!`09Requires MX logicals to be defined X$!`09Requires MX software to be present X$! X$! Revision History: X$!`091.0`09??????????`09Hunter Goatley`09Original version X$!`091.1`091-NOV-1993`09Andy Harper`09Added parameters, comments, error checki Vng X$! X$ X$ On control_y then $ goto RestartMX X$ set noon X$ em = "write sys$output ""%"+f$parse(f$environment("PROCEDURE"),,,"NAME")+""" V," X$ mcp = "$mx_exe:mcp.exe/file=mx_dir:mx_config.mxcfg" X$ MXstartproc = "sys$startup:mx_\0B07" X$ X$ X$! --------\0808\1010\2020\400C X$! Set up the parameters; default if necessary X$! --------\0808\1010\2020\400C X$ Delay = P1 X$ if Delay .eqs. "" then $ \1909= "00:00:15"`09! Interval between tries X$ MaxTries = P2 X$ if maxTries .eqs. "" then $ M\1908= 240`09`09! Number of retries; an hour ma Vximum X$ Username = P3 X$ if Username .eqs. "" then $ \1909= "SYSTEM"`09!\2F0Afor MX startup X$ X$ X$! --------\0808\1010\2020\400C X$! Check privileges required X$! --------\0808\1010\2020\400C X$ if .not. f$priviledge("SYSPRV,CMKRNL") X$ then X$ em "-E-INSUFPRV, Insufficient privilege" X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Make sure we have access to the MCP program so we can shut down everything V X$! --------\0808\1010\2020\400C X$ if f$search("mx_exe:mcp.exe") .eqs. "" X$ then X$ em "-E-NOMCP, MX control program MCP not found" X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Shut all the MX processes having the queue file open X$! --------\0808\1010\2020\400C X$ em "-I-SHUTDOWN, Shutting down MX" X$ mcp shutdown X$ X$ X$! --------\0808\1010\2020\400C X$! Loop until we can OPEN the SYSTEM_QUEUE file; when \2806, nobody X$! else is using it, which means all the agents have shutdown. X$! However, to avoid MX being shutdown forever due to any system problems X$! elsewhere, limit the amount of time we retry (see P1 and p2). X$! --------\0808\1010\2020\400C X$ Tries = 0 X$ wait_loop: X$ if Tries .gt. Max\0E06then $ goto Wait_Too_Long X$ Tries =\0807+ 1 X$ wait 'Delay' X$ open/read/error=wait_loop tmp mx_flq_dir:system_queue.flq_ctl X$ close tmp X$ X$! --------\0808\1010\2020\400C X$! Now we can do the compress;\0A09 down to a temporary file first and X$! then rename it to the master one. This prevents unexpected aborts from X$! leaving partial queue files. X$! X$! NOTE: The FDL file used here should be tailored to the site by modifying X$! the ALLOCATION and EXTENSION values in AREA 0 and AREA 1. To avoid X$! excessive extent pointers building up, it is suggested that the X$! ALLOCATION quantity should be close to the maximum size reached by X$! the queue file. Excessive extent pointers will cause a serious X$! performance hit. X$! --------\0808\1010\2020\400C X$ em "-I-COMPRESS, Compressing MX queue file" X$ convert/fdl=mx_flq_dir:system_queue \1817.flq_ctl - X mx_flq_dir:system_queue.flq_ctl_tmp; X$ if .not. $status X$ then X$ em "-I-STATUS, ", $status, " : ", f$message(\1A07) X$ em "-I-COMPFAIL, Unable to complete the queue compress - aborted" X$ delete mx_flq_dir:system_queue.flq_ctl_tmp;* X$ goto CompressNotDone X$ endif X$ X$! --------\0808\1010\2020\400C X$! Rename the newly built queue file to the master\1306and purge off old X$! copies X$! --------\0808\1010\2020\400C X$ em "-I-UPDATE, Updating queue file" X$ rename mx_flq_dir:system_queue.flq_ctl_tmp \0D08 X$ purge mx_flq_dir:system_queue.flq_ctl X$ rename mx_flq_dir:system_queue.flq_ctl; ;1 X$ X$ X$! --------\0808\1010\2020\400C X$! Compress aborted, finished or timed out; restart MX X$! --------\0808\1010\2020\400C X$CompressNotDone: X$Wait_Too_Long: X$RestartMx: X$ em "-I-RESTART, Restarting MX" X$ if f$edit(f$getjpi("","USERNAME"),"UPCASE,TRIM") .eqs. Username X$ then X$ @'MXstartproc' X$ else X$ submit/user='Username' 'MXstartproc' X$ endif $ call unpack MX_SHRINK.COM;9 1610996648 "" 10 1 2 $! $ create 'f' XMX QUEUE FILE COMPRESSION X X3-NOV-1993 X XAndy Harper X X--------\0808\1010\2020\400F X X XINTRODUCTION X XThe queue file used by the MX mail system (MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL) Xwill, on a heavily used system, grow unbounded. This causes performance Xproblems due to excessive use of file extent pointers (hence lots of disk Xseeks). X XThe DCL procedure described here is an attempt to reduce the performance Xproblems by periodically compressing the queue file. X X XOPERATION X XThe queue file cannot be compressed while MX is running (strictly,\1F07any Xuser has the file open) so MX needs to be shut down. The procedure will first V Xshut MX down, then compress the queue file, and finally restart MX. Depending V Xon the size of the queue file, MX may be unavailable for some considerable Xtime. It is recommended that this procedure be run at the quietest moment Xovernight. X XThe procedure attempts to be fail safe. When the shutdown signal has been sen Vt Xto MX, it loops trying to open the queue file. Once it can, no-one is accessin Vg Xit and this indicates that MX has completely shut down (after finishing Xwhatever it was doing in a tidy manner). However, should the procedure be Xunable to open the queue file after a number of retries, it will give up and Xrestart MX. X XMX is also restarted if the procedure is aborted or\1F08compression attempt Xfails in some way. X XOnce the queue is accessible to the procedure, the compression operation Xbegins. A new file is created using the currently valid records of the existin Vg Xqueue. This new file is given a different name (and later renamed) to avoid a V Xpartial queue file being inadvertently put in place as the master. X XThe FDL file supplied with MX is used to create the queue\3206and it is Xstrongly recommended that the ALLOCATION and EXTENSION parameters of this file V Xbe tailored to the site; the exact values will depend on how much the queue Xfile grows during a typical day. It is recommended that the ALLOCATION Xparameters be set to regenerate the file close to its maximum size. X X XRUNNING THE PROCEDURE X XThe procedure is run as follows: X X @MX_SHRINK `5Bretry`5D `5Bdelay`5D `5Busername`5D X Xwhere: X X `5Bretry`5D`09is the number of times to retry`09if the queue file cannot be V X`09`09opened (default: 240) X X `5Bdelay`5D`09is the delta time to wait between retries `5Bdefault: 15 seco Vnds`5D X X `5Busername`5D`09is the name under which to restart MX `5Bdefault: SYSTEM V`5D X X XYou should arrange to run this procedure on a regular basis every night by som Ve Xexternal means. X X X XACKNOWLEDGEMENTS X XI'm grateful to Hunter Goatley, who provided the basic draft of this procedure V, Xfor permission to make his code freely available in my modified form. $ call unpack MX_SHRINK.DOC;2 2114130419 "" 6 2 2 $ v=f$verify(v) $ exit ================================================================================ Archive-Date: Thu, 04 Nov 1993 15:32:11 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 16:26:26 EST From: "Steve Thompson, Cheme System Mangler" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU CC: olin@cheme.cornell.edu Message-ID: <009750B7.CF2CB0F4.7274@cheme.cornell.edu> Subject: RE: Compressing the MX queue Andy Harper writes: >Following recent discussions on keeping the MX queue file clean, I've enclosed >a short DCL procedure which will do the job. It's based on Hunter Goatley's >original but I've added some improvements, such as error checking and making it >fail safe. After a quick scan of the procedure, it appears to be useable only if MX is running on a single machine in a cluster. Firstly, "mcp shutdown" doesn't shut down MX processes on other cluster nodes (at least for me). This one is easily fixed (add /cluster). Secondly, it submits the MX restart job only on the node on which it is running; the queue used is SYS$BATCH. The submit code needs to be replicated for each cluster member. Given this, looks neat! -steve --------------------------------------------------------------------------- Steve Thompson, System Mangler Internet: thompson@cheme.cornell.edu School of Chemical Engineering Bitnet: thompson@crnlchme Olin Hall, Cornell University Phone: (607) 255 5573 Ithaca NY 14853 FAX: (607) 255 9166 --------------------------------------------------------------------------- ================================================================================ Archive-Date: Thu, 04 Nov 1993 15:41:17 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 15:34:45 CST From: "George D. Greenwade" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU CC: udaa055@bay.cc.kcl.ac.uk Message-ID: <009750B0.9697B560.28919@SHSU.edu> Subject: RE: Compressing the MX queue On Thu, 04 Nov 1993 16:40:49 EST, Andy Harper posted: > Following recent discussions on keeping the MX queue file clean, I've > enclosed a short DCL procedure which will do the job. It's based on Hunter > Goatley's original but I've added some improvements, such as error checking > and making it fail safe. > > It's designed to be run overnight (via whatever mechanism you use to auto > run such things - DECscheduler, CRON etc.) and will shutdown MX, compress > the file and then restart MX when done using the new shortened file. It > will automatically restart MX if aborted partway, if the compression fails > for any reason, or if it can't get exclusive access to the queue file > within a reasonable time (configurable). > > I'm open to suggestions for improvements. I haven't VMS_SHARE'd the files, but they look close to what we do here. I did note one *critical* factor which you ought to consider in this: > X$! --------\0808\1010\2020\400C > X$! Shut all the MX processes having the queue file open > X$! --------\0808\1010\2020\400C > X$ em "-I-SHUTDOWN, Shutting down MX" > X$ mcp shutdown That last line really ought to be > X$ mcp shutdown/cluster ^^^^^^^^ as this is to be run one one node; if you have multiple nodes involved, the "mcp shutdown" won't get everything and mx_flq_dir:system_queue.flq_ctl will remain open. Obviously, don't edit the VMS_SHARE'd file (don't have to tell you that, do I, Andy? 8-)), but edit it once extracted. --George %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% George D. Greenwade, Ph.D. Bitnet: BED_GDG@SHSU Department of Economics and Business Analysis THEnet: SHSU::BED_GDG College of Business Administration Voice: (409) 294-1266 P. O. Box 2118 FAX: (409) 294-3612 Sam Houston State University Internet: bed_gdg@SHSU.edu Huntsville, TX 77341 bed_gdg%SHSU.decnet@relay.the.net %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ================================================================================ Archive-Date: Thu, 04 Nov 1993 16:00:01 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 14:54:13 MST From: "Michael L. Hitch" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <009750AA.ED0502A0.30093@msu.oscs.montana.edu> Subject: RE: Compressing the MX queue "Steve Thompson, Cheme System Mangler" writes: > After a quick scan of the procedure, it appears to be useable only if > MX is running on a single machine in a cluster. Firstly, "mcp shutdown" > doesn't shut down MX processes on other cluster nodes (at least for > me). This one is easily fixed (add /cluster). Secondly, it submits > the MX restart job only on the node on which it is running; the queue > used is SYS$BATCH. The submit code needs to be replicated for each > cluster member. Given this, looks neat! I've been working on a similar procedure. I am using SYSMAN to execute the startup procedure on all nodes of the cluster. Michael --- Michael L. Hitch osymh@msu.oscs.montana.edu Computer Consultant OSYMH@MTSUNIX1.BITNET Office of Systems and Computing Services Montana State University Bozeman, MT USA ================================================================================ Archive-Date: Thu, 04 Nov 1993 16:29:26 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 16:24:23 CDT From: "John Clement-Rice U. (713)527-8101x2037" Reply-To: MX-List@WKUVX1.BITNET To: MX-LIST@WKUVX1.BITNET CC: clement@physics.rice.edu Message-ID: <009750B7.85C7A540.17885@physics.rice.edu> Subject: How to make a comment ?? Whenever I send mail to a "listserver" it bounces my signature. How do I make it ignore the signature, or turn it into a comment. See below: //\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ //John M. Clement Hepnet: UTSPAN::RIPHYS::CLEMENT //Bonner Lab Bitnet: CLEMENT@RICEVM1 //Rice University Internet: CLEMENT@PHYSICS.RICE.EDU //PO Box 1892 BellTel: (713)527-8181 x2037 //Houston, Tx. 77251 Fax: (713)285-5215 //\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ ================================================================================ Archive-Date: Thu, 04 Nov 1993 16:42:57 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 16:42:28 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET CC: clement@physics.rice.edu Message-ID: <009750BA.0C5A9E03.15983@WKUVX1.BITNET> Subject: RE: How to make a comment ?? "John Clement-Rice U. (713)527-8101x2037" writes: > >Whenever I send mail to a "listserver" it bounces my signature. >How do I make it ignore the signature, or turn it into a comment. >See below: > Add the QUIT command after the list server commands: ADD xxxxx ADD xxxxx QUIT signature Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Thu, 04 Nov 1993 16:45:13 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 4 Nov 1993 16:43:46 -0600 (CST) From: "H.A. Kippenhan Jr." Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <931104164346.2a801c2a@fndcd.fnal.gov> Subject: RE: How to make a comment ?? Hi John: | | From: SMTP%"MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU" 4-NOV-1993 16:31:55.43 | Subj: How to make a comment ?? | | Whenever I send mail to a "listserver" it bounces my signature. | How do I make it ignore the signature, or turn it into a comment. | See below: | This is also true of the "-request" address for any E-mail list supported by MX. I just do a send/noed from the VMS mail prompt so I don't tack my sig. file on the end of the few things I send to mail list servers. Regards - Kipp - {~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~} { H.A. Kippenhan Jr. | Internet: Kippenhan@FNDCD.FNAL.GOV } { National HEPnet Management | HEPnet/NSI DECnet: FNDCD::KIPPENHAN } { Fermi National Accelerator Lab. | BITnet: Kippenhan@FNDCD.BITNET } { P.O. Box 500 MS: FCC-3E/368 | Telephone: (708) 840-8068 } { Batavia, Illinois 60510 | FAX: (708) 840-8463 } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ================================================================================ Archive-Date: Thu, 04 Nov 1993 16:54:13 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 16:41:45 CST From: Kenny Kon Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <009750B9.F2A54BC0.1453@bible.acu.edu> Subject: RE: How to make a comment ?? Date sent: 4-NOV-1993 16:41:08 >Whenever I send mail to a "listserver" it bounces my signature. >How do I make it ignore the signature, or turn it into a comment. >See below: I think a /QUIT after your command will do it. --------------------------------------------------------------------------- Kenny Kon Internet: kon@bible.acu.edu ACU Station, Box 7725 kkon@acuvax.acu.edu Abilene, TX 79699 BITNET: kkon@acuvax.bitnet (915)672-4915 CompuServe: 72724.3547@compuserve.com (915)674-3708 (work) ================================================================================ Archive-Date: Thu, 04 Nov 1993 16:55:58 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 16:49:19 CST From: "George D. Greenwade" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU CC: olin@cheme.cornell.edu Message-ID: <009750BB.01759AA0.29324@SHSU.edu> Subject: RE: Compressing the MX queue On Thu, 04 Nov 1993 16:26:26 EST, Steve Thompson posted: > After a quick scan of the procedure, it appears to be useable only if MX is > running on a single machine in a cluster. Firstly, "mcp shutdown" doesn't > shut down MX processes on other cluster nodes (at least for me). This one > is easily fixed (add /cluster). Secondly, it submits the MX restart job > only on the node on which it is running; the queue used is SYS$BATCH. The > submit code needs to be replicated for each cluster member. Given this, > looks neat! There shouldn't be any reason to submit it but on the one node which is going to do the cleanup (the fastest, I would hope) if you are using a cluster setup (you've only got to worry about one MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL file within the cluster, right?). One trick we added was to not use SYS$BATCH for this type of work; instead we created a different batch queue, MX$nodename_BATCH for everything MX-related which is only used by known ACL-restricted MX-related users. This way, MX events aren't stuffed behind a lot of other files in SYS$BATCH (we do an enormous amount of overnight scheduling on batches; but I'm sure most everyone does) and can better control the timing of MX-related events. Regards, George %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% George D. Greenwade, Ph.D. Bitnet: BED_GDG@SHSU Department of Economics and Business Analysis THEnet: SHSU::BED_GDG College of Business Administration Voice: (409) 294-1266 P. O. Box 2118 FAX: (409) 294-3612 Sam Houston State University Internet: bed_gdg@SHSU.edu Huntsville, TX 77341 bed_gdg%SHSU.decnet@relay.the.net %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ================================================================================ Archive-Date: Thu, 04 Nov 1993 17:41:07 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: Compressing the MX queue Message-ID: <1993Nov4.152057.271@buckie.hsc.colorado.edu> From: Date: 4 Nov 93 15:20:57 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17183879@MVB.SAIC.COM>, "Andy, Systems Manager" writes: >Following recent discussions on keeping the MX queue file clean, I've enclosed >a short DCL procedure which will do the job. It's based on Hunter Goatley's >original but I've added some improvements, such as error checking and making it >fail safe. Looks great. I agree with Steve Thompson's comment that MX needs to be stopped and re-started cluster-wide. I'd recommend using SYSMAN to submit to the SYS$BATCH queues on all nodes (requires OPER), but this doesn't ensure that SYS$BATCH is not a generic queue for the cluster, though... (My little snippet below assumes there is a nodename_SYSMGR$BATCH queue on all nodes in the cluster.) SYSLCK is required for the MCP SHUTDOWN command (see MX_DOC:MGMT_GUIDE in MX V3.3). If you use SYSMAN, be sure to override any SYSMANINI logical names, and use a TIMEOUT (in case there's a system shutdown or reboot going on in the middle of the night): ... $ $ DEFINE/USER_MODE SYSMANINI NLA0: $ RUN SYS$SYSTEM:SYSMAN SET TIMEOUT 00:00:30 SET ENVIRONMENT/CLUSTER DO SUBMIT/USER=SYSTEM/QUEUE='F$GETSYI("NODENAME")'_SYSMGR$BATCH - SYS$STARTUP:MX_STARTUP.COM EXIT $ ... -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Thu, 04 Nov 1993 18:03:36 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: How to make a comment ?? Message-ID: <1993Nov4.160646.273@buckie.hsc.colorado.edu> From: Date: 4 Nov 93 16:06:46 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17186049@MVB.SAIC.COM>, "John Clement-Rice U. (713)527-8101x2037" writes: >Whenever I send mail to a "listserver" it bounces my signature. >How do I make it ignore the signature, or turn it into a comment. >See below: (You failed to attach an example, but I think I understand the problem). "Most" LISTSERVs support a QUIT command, which you can put as the last command you're sending to the LISTSERV. MX's "MXserver" supports QUIT; I'm not sure about Eric Thomas's LISTSERV (I couldn't find a reference to QUIT in the short HELP file I just retrieved from UGA.BITNET). The best solution, of course, is to not append your signature to the mail message in the first place! If you're using MX (I hope you are, because you posted your question to the MX list/newsgroup), you should deassign the logical name MX_AUTO_SIGNATURE to prevent MX from automatically adding your signature to your message (see section 1.4.1 of the MX V3.3 user's guide for information on this). -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Thu, 04 Nov 1993 18:57:06 CST Sender: list-mgr@WKUVX1.BITNET Subject: RE: Compressing the MX queue Message-ID: <1993Nov4.163153.275@buckie.hsc.colorado.edu> From: Date: 4 Nov 93 16:31:53 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17186957@MVB.SAIC.COM>, "George D. Greenwade" writes: >On Thu, 04 Nov 1993 16:26:26 EST, Steve Thompson >posted: >> After a quick scan of the procedure, it appears to be useable only if MX is >> running on a single machine in a cluster. Firstly, "mcp shutdown" doesn't >> shut down MX processes on other cluster nodes (at least for me). This one >> is easily fixed (add /cluster). Secondly, it submits the MX restart job >> only on the node on which it is running; the queue used is SYS$BATCH. The >> submit code needs to be replicated for each cluster member. Given this, >> looks neat! > >There shouldn't be any reason to submit it but on the one node which is >going to do the cleanup (the fastest, I would hope) if you are using a >cluster setup (you've only got to worry about one >MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL file within the cluster, right?). Right, but you need to shutdown the router, local agent, incoming SMTP agents, outgoing SMTP agents, etc., on all nodes so none of them try to open the file (or you won't be able to open it to tune it). Then you need to start them up again when you're done tuning.... -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Fri, 05 Nov 1993 05:55:19 CST Sender: list-mgr@WKUVX1.BITNET Date: Thu, 04 Nov 1993 23:14:06 CST From: Kenny Kon Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <009750F0.C1FD86E0.1553@bible.acu.edu> Subject: Using ADD Date sent: 4-NOV-1993 23:11:57 I have some problems using the ADD command. Seems like people I add manually have problems posting to the list. Whereas those that subscribe have no problems. However, those that are ADDed have no problems receiving posts. I have an old mailing list from someone else and I ADDed these on manually. --------------------------------------------------------------------------- Kenny Kon Internet: kon@bible.acu.edu ACU Station, Box 7725 kkon@acuvax.acu.edu Abilene, TX 79699 BITNET: kkon@acuvax.bitnet (915)672-4915 CompuServe: 72724.3547@compuserve.com (915)674-3708 (work) ================================================================================ Archive-Date: Fri, 05 Nov 1993 06:08:53 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 06:08:27 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <0097512A.A4A037C5.16251@WKUVX1.BITNET> Subject: RE: Using ADD Kenny Kon writes: > >I have some problems using the ADD command. Seems like people I add >manually have problems posting to the list. Whereas those that subscribe >have no problems. However, those that are ADDed have no problems >receiving posts. I have an old mailing list from someone else and I ADDed >these on manually. > MX maintains the case of the user portion of an address. If you've got the list protected so that only subscribers can post, then the case of their list entry *has* to match the case of their incoming mail. For example, if I was on the list as GOATHUNTER@WKUVX1.BITNET but my mail arrived to you as goathunter@WKUVX1.BITNET, I would not be allowed to post. If the list is large, you could ADD them all /NOCASE, which tells MX to ignore case when making protection checks. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Fri, 05 Nov 1993 07:53:44 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 08:31:09 EST From: "Brian Tillman" Reply-To: MX-List@WKUVX1.BITNET To: WKUVX1.BITNET!mx-list@esseye.si.com Message-ID: <0097513E.93CE44A0.10617@swdev.si.com> Subject: Re: How to make a comment ?? In article <17186049@MVB.SAIC.COM>, "John Clement-Rice U. (713)527-8101x2037" writes: >Whenever I send mail to a "listserver" it bounces my signature. >How do I make it ignore the signature, or turn it into a comment. >See below: So far, no one's mentioned the original method of suppressing an auto-signature: Simply put "/nosig" in your message. -----------------------------+-------------------------------- Brian Tillman | Internet: tillman@swdev.si.com Smiths Industries, Inc. | tillman_brian@si.com 4141 Eastern Ave., MS129 | Hey, I said this stuff myself. Grand Rapids, MI 49518-8727 | My company has no part in it. -----------------------------+-------------------------------- ================================================================================ Archive-Date: Fri, 05 Nov 1993 08:56:52 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: How to make a comment ?? Message-ID: <2bdi50$akc@umd5.umd.edu> From: Date: 5 Nov 1993 12:49:36 GMT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <1993Nov4.160646.273@buckie.hsc.colorado.edu>, dwing@uh01.Colorado.EDU (Dan Wing) writes: >In article <17186049@MVB.SAIC.COM>, "John Clement-Rice U. (713)527-8101x2037" writes: > >>Whenever I send mail to a "listserver" it bounces my signature. >>How do I make it ignore the signature, or turn it into a comment. ... >The best solution, of course, is to not append your signature to the >mail message in the first place! ... >deassign the logical name MX_AUTO_SIGNATURE to prevent MX from >automatically adding your signature to your message (see section 1.4.1 of >the MX V3.3 user's guide for information on this). It should not be necessary to deassign. I just have /NOSIG (case insensitive) as the last line of the message. >-Dan Wing, Systems Administrator, University Hospital, Denver > dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Fri, 05 Nov 1993 09:17:52 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 16:13:37 EST From: figon@OBMARA.CNRS-MRS.FR Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET CC: figon@OBMARA.CNRS-MRS.FR Message-ID: <0097517F.2EE489E6.23355@OBMARA.CNRS-MRS.FR> Subject: problem receiving mail on AXP/VMS system. On a AXP/openVMS V1.5, UCX 3.0 and MX 3.3 system, sending mail works ok, but it is not possible to receive mail. "Postmaste" returns the following error message : Is this a known problem ? Thank you, Patrick Figon Observatoire de Marseille France *********************************************************************** From: MX%"Postmaster@OBMARA.CNRS-MRS.FR" 3-NOV-1993 22:14:56.42 To: FIGON CC: Subj: SMTP delivery error Return-Path: Received: from ALPHAB by OBMARA (MX V3.3 VAX) with SMTP; Wed, 03 Nov 1993 22:14:54 EST Date: Wed, 03 Nov 1993 22:14:57 EST X-MX-Warning: Warning -- Invalid "From" header. From: Wed, 03 Nov 1993 22:14:57 EST To: Subject: SMTP delivery error Note: this message was generated automatically. A problem occurred during SMTP delivery of your message. Error occurred sending to the following user(s): (via alphab.cnrs-mrs.fr): %MX_SMTP-F-TRANSACTION_FAI, transaction failed Transcript: Rcvd: 220 ALPHAB MX V3.3 AXP SMTP server ready at Wed, 03 Nov 1993 22:14:56 EST Sent: HELO ALPHAB Rcvd: 250 Hello, ALPHAB Sent: MAIL FROM: Rcvd: 250 MAIL command accepted. Sent: RCPT TO: Rcvd: 250 Recipient okay (at least in form) Sent: DATA Rcvd: 354 Start mail input; end with . Rcvd: 554 Received too many times by this host. Sent: QUIT Rcvd: 221 ALPHAB Service closing transmission channel ======================================================================== Message follows. Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:55 EST Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:54 EST Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:52 EST Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:50 EST Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:48 EST Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:47 EST Received: from ALPHAB by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:45 EST Received: from OBMARA by ALPHAB (MX V3.3 AXP) with SMTP; Wed, 03 Nov 1993 22:14:44 EST Received: by OBMARA.CNRS-MRS.FR (MX V3.3 VAX) id 23049; Wed, 03 Nov 1993 22:14:37 EST Date: Wed, 03 Nov 1993 22:14:37 EST From: figon@OBMARA.CNRS-MRS.FR To: system@alphab.cnrs-mrs.fr CC: figon@OBMARA.CNRS-MRS.FR Message-ID: <0097501F.48596326.23049@OBMARA.CNRS-MRS.FR> Subject: test sss ================================================================================ Archive-Date: Fri, 05 Nov 1993 09:30:24 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 09:29:45 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <00975146.C3BF8072.16423@WKUVX1.BITNET> Subject: RE: problem receiving mail on AXP/VMS system. figon@OBMARA.CNRS-MRS.FR writes: > > On a AXP/openVMS V1.5, UCX 3.0 and MX 3.3 system, sending mail works >ok, but it is not possible to receive mail. "Postmaste" returns the >following error message : > Is this a known problem ? Sounds like you need to follow the instructions for UCX in the MX V3.3 release notes. Included below. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ------------------------------------------------------------------------------- 1.3 VMS/ULTRIX Connection Notes If you are using DEC VMS/ULTRIX Connection (UCX) V1.2 through V2.0, you should review your UCX configuration to ensure that you are using your fully-qualified domain name (FQDN) as your primary local host name in your UCX host table. If you followed the UCX documentation when configuring UCX with UCX$CONFIG, it is highly likely that you did not specify your FQDN as your primary host name. If your abbreviated host name is listed as the primary in your UCX host table, then messages sent from the local system will go out with an unreplyable return address. To check your local host name, use the UCX utility: $ UCX UCX> SHOW HOST/LOCAL x.x.x.x ! use your IP address here To ensure that your FQDN is the primary host name for your address, use the UCX commands: UCX> SET NOHOST x.x.x.x ! use your IP address here UCX> SET HOST "full.qual.dom.name"/ADDR=x.x.x.x/ALIAS="abbrev" You will be asked for confirmation on the SET NOHOST command. Substitute your FQDN and IP address in the SET HOST command, and if you would still have your shortened host name in the host table, use the /ALIAS qualifier, as shown. Once you have updated your host table, you should review the file SYS$MANAGER:UCX$INET_SET_ INTERFACES.COM and replace all instances of your abbreviated host name with your FQDN. Note: Remember that UCX is case sensitive with regard to host names! ================================================================================ Archive-Date: Fri, 05 Nov 1993 09:58:04 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: problem receiving mail on AXP/VMS system. Message-ID: <1993Nov5.083651.281@buckie.hsc.colorado.edu> From: Date: 5 Nov 93 08:36:50 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17215281@MVB.SAIC.COM>, figon@OBMARA.CNRS-MRS.FR writes: > On a AXP/openVMS V1.5, UCX 3.0 and MX 3.3 system, sending mail works >ok, but it is not possible to receive mail. "Postmaste" returns the >following error message : > Is this a known problem ? [...] >X-MX-Warning: Warning -- Invalid "From" header. >From: Wed, 03 Nov 1993 22:14:57 EST [...] > Transcript: > Rcvd: 220 ALPHAB MX V3.3 AXP SMTP server ready at Wed, 03 Nov 1993 22:14:56 >EST > Sent: HELO ALPHAB > Rcvd: 250 Hello, ALPHAB > Sent: MAIL FROM: > Rcvd: 250 MAIL command accepted. > Sent: RCPT TO: > Rcvd: 250 Recipient okay (at least in form) > Sent: DATA > Rcvd: 354 Start mail input; end with . > Rcvd: 554 Received too many times by this host. > Sent: QUIT It looks like ALPHAB isn't defined as a LOCAL delivery in your PATH settings. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Fri, 05 Nov 1993 11:52:08 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: Disabling SMTP Message-ID: <2bdtab$s30@gap.cco.caltech.edu> From: Date: 5 Nov 1993 16:00:11 GMT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17181179@MVB.SAIC.COM>, Kenny Kon writes: =Date sent: 3-NOV-1993 23:30:55 = =Is it normal to be able to use SMTP%"address" after installing MX? Now all =the mail comes in via MX, but mail can still be sent via MX%"address" or =SMTP%"address" Could someone tell me how to disable SMTP so that everyone =will be able to use only MX%. Somewhere in the MX manual it tells you how to do this. One way is: $ DEFINE/SYS/EXEC SMTP_MAILSHR MX_MAILSHR -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Fri, 05 Nov 1993 11:52:26 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: How to make a comment ?? Message-ID: <1993Nov5.091445.497@tachyon.com> From: Reply-To: MX-List@WKUVX1.BITNET Date: 5 Nov 93 09:14:45 CST To: MX-List@WKUVX1.BITNET In article <1993Nov4.160646.273@buckie.hsc.colorado.edu>, dwing@uh01.Colorado.EDU (Dan Wing) writes: > In article <17186049@MVB.SAIC.COM>, "John Clement-Rice U. (713)527-8101x2037" writes: > >>Whenever I send mail to a "listserver" it bounces my signature. >>How do I make it ignore the signature, or turn it into a comment. >>See below: > [stuff about "quit" deleted] > > The best solution, of course, is to not append your signature to the > mail message in the first place! If you're using MX (I hope you are, > because you posted your question to the MX list/newsgroup), you should > deassign the logical name MX_AUTO_SIGNATURE to prevent MX from > automatically adding your signature to your message (see section 1.4.1 of > the MX V3.3 user's guide for information on this). I prefer using /nosig in the message itself. Auto signature is too useful to disable. > -Dan Wing, Systems Administrator, University Hospital, Denver Wayne -- ============================================================================== Wayne Sewell |INET: wayne@tachyon.com Tachyon Software Consulting |UUCP: uupsi!uupsi6!tachyon!wayne P. O. Box 550937, Dallas TX 75355-0937 |Voice: (214)-553-9760, Fax: -553-0077 ============================================================================= Cute Girl, to Curly:"Oh, what a beautiful head of bone you have!" ================================================================================ Archive-Date: Fri, 05 Nov 1993 12:58:58 CST Sender: list-mgr@WKUVX1.BITNET Date: 5 Nov 93 12:40:00 CST From: "John Clement-Rice U. (713)527-8101x2037" Reply-To: MX-List@WKUVX1.BITNET Subject: Re: How to make a comment ?? To: "MX-List" Thanks for all the replies, I could just put QUIT in my signature, and then forget about the whole thing. John Clement ================================================================================ Archive-Date: Fri, 05 Nov 1993 13:09:14 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 18:36:48 EST From: "Andy, Systems Manager" Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET CC: udaa055@BAY.CC.KCL.AC.UK Message-ID: <00975193.2F7750A0.7036@bay.cc.kcl.ac.uk> Subject: MX Queue compression - update Dear All, Thanks for the various suggestions to the MX_SHRINK procedure recently posted. As a result, I've updated the code, which is enclosed below. Most of you wanted it to work cluster-wide; I hope this bit now works. Unfortunately, I no longer have access to a real cluster on which to try it so please let me know if anything falls apart. The biggest difficulty with clustering, as others have pointed out, are restarting the MX code correctly on all nodes after compression has taken place and dealing with clusters that have multiple queue files, perhaps on different system disks. I believe the first problem is solved (see docs + code for details). The second problem has not been tackled yet - any suggestions as to how best to do it gratefully received. Regards, Andy Harper Kings College London $! ------------------ CUT HERE ----------------------- $ v='f$verify(f$trnlnm("SHARE_UNPACK_VERIFY"))' $! $! This archive created: $! Name : MX_SHRINK $! By : Andy Harper $! Date : 5-NOV-1993 18:28:17.74 $! Using: VMS_SHARE 8.4, (C) 1993 Andy Harper, Kings College London UK $! $! Credit is due to these people for their original ideas: $! James Gray, Michael Bednarek $! $! TO UNPACK THIS SHARE FILE, CONCATENATE ALL PARTS IN ORDER $! AND EXECUTE AS A COMMAND PROCEDURE ( @name ) $! $! THE FOLLOWING FILE(S) WILL BE CREATED AFTER UNPACKING: $! 1. MX_SHRINK.COM;10 $! 2. MX_SHRINK.DOC;3 $! $ set="set" $ set symbol/scope=(nolocal,noglobal) $ f=f$parse("SHARE_UNPACK_TEMP","SYS$SCRATCH:."+f$getjpi("","PID")) $ e="write sys$error ""%UNPACK"", " $ w="write sys$output ""%UNPACK"", " $ if .not. f$trnlnm("SHARE_UNPACK_LOG") then $ w = "!" $ if f$getsyi("CPU") .gt. 127 then $ goto start $ ve=f$getsyi("version") $ if ve-f$extract(0,1,ve) .ges. "4.4" then $ goto start $ e "-E-OLDVER, Must run at least VMS 4.4" $ v=f$verify(v) $ exit 44 $unpack: subroutine ! P1=filename, P2=checksum, P3=attributes,P4=size $ if f$parse(P1) .nes. "" then $ goto dirok $ dn=f$parse(P1,,,"DIRECTORY") $ w "-I-CREDIR, Creating directory ''dn'" $ create/dir 'dn' $ if $status then $ goto dirok $ e "-E-CREDIRFAIL, Unable to create ''dn' File skipped" $ delete 'f'* $ exit $dirok: $ x=f$search(P1) $ if x .eqs. "" then $ goto file_absent $ e "-W-EXISTS, File ''P1' exists. Skipped" $ delete 'f'* $ exit $file_absent: $ w "-I-UNPACK, Unpacking ", P5, " of ", P6, " - ", P1, " - ", P4, " Blocks" $ n=P1 $ if P3 .nes. "" then $ n=f $ if .not. f$verify() then $ define/user sys$output nl: $ EDIT/TPU/NOSEC/NODIS/COM=SYS$INPUT/NOJOURNAL 'f'/OUT='n' PROCEDURE GetHex(s,p)LOCAL x1,x2;x1:=INDEX(t,SUBSTR(s,p,1))-1;x2:=INDEX(t, SUBSTR(s,p+1,1))-1;RETURN 16*x1+x2;ENDPROCEDURE;PROCEDURE SkipPartsep LOCAL m; LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m);EXITIF INDEX(ERASE_LINE, "-+-+-+-+-+-+-+-+")=1;ENDLOOP;ENDPROCEDURE;PROCEDURE ProcessLine LOCAL c,s,l,b, n,p;c := ERASE_CHARACTER(1);s := ERASE_LINE;EDIT(s,"TRIM");IF c = "X" THEN SPLIT_LINE; ENDIF;MOVE_HORIZONTAL(-1);l := LENGTH(s);p := 1; LOOP EXITIF p > l;c := SUBSTR(s,p,1);p := p+1;CASE c FROM ' ' TO '`' ['\']: b:= GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT( SUBSTR(CURRENT_LINE, CURRENT_OFFSET-b+1,n));['&']: b:=GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT(ASCII(n)*b);['`']: COPY_TEXT(ASCII(GetHex(s,p))); p:=p+2;[INRANGE, OUTRANGE]: COPY_TEXT(c);ENDCASE;ENDLOOP;ENDPROCEDURE;PROCEDURE Decode LOCAL m; POSITION(BEGINNING_OF(b));LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m); IF INDEX(CURRENT_LINE,"+-+-+-+-+-+-+-+-")=1 THEN SkipPartSep;ELSE ProcessLine; MOVE_HORIZONTAL(1);ENDIF;ENDLOOP;ENDPROCEDURE;SET(FACILITY_NAME,"UNPACK");SET( SUCCESS,OFF);SET(INFORMATIONAL,OFF);t:="0123456789ABCDEF";f:=GET_INFO( COMMAND_LINE,"file_name");b:=CREATE_BUFFER(f,f);Decode;WRITE_FILE(b,GET_INFO( COMMAND_LINE,"output_file"));QUIT; $ if p3 .eqs. "" then $ goto dl $ open/write fdl &f $ write fdl "RECORD" $ write fdl P3 $ close fdl $ w "-I-CONVRFM, Converting record format to ", P3 $ convert/fdl=&f &f-1 &P1 $dl: delete 'f'* $ checksum 'P1' $ if checksum$checksum .nes. P2 then $ - e "-E-CHKSMFAIL, Checksum of ''P1' failed." $ exit $ endsubroutine $start: $! $ create 'f' X$! Purpose: X$!`09Automatically compress the MX QUEUE file X$! X$! Parameters: X$!`09P1`09List of queues for restart job. if a stand-alone node, this X$!`09`09is optional and defaults to SYS$BATCH. If a member of a cluster X$!`09`09a list of execution batch queues (one per node) is required X$!`09P2`09`5Boptional`5D delay time between queue file access retries (default V: 15 seconds) X$!`09P3`09`5Boptional`5D number of retries (default: 240 ) X$!`09P4`09`5Boptional`5D username to run MX (default: SYSTEM) X$! X$! Privileges: X$!`09SYSPRV X$!`09CMKRNL X$!`09SYSLCK X$!`09 X$! Environment: X$!`09Requires MX logicals to be defined X$!`09Requires MX software to be present X$! X$! Revision History: X$!`091.0`09??????????`09Hunter Goatley`09Original version X$!`091.1`091-NOV-1993`09Andy Harper`09Added parameters, comments, error checki Vng X$!`091.2`095-NOV-1993`09Andy Harper`09Add cluster support; fix privs X$! X$ X$ On control_y then $ goto RestartMX X$ set noon X$ em = "write sys$output ""%"+f$parse(f$environment("PROCEDURE"),,,"NAME")+""" V," X$ mcp = "$mx_exe:mcp.exe/file=mx_dir:mx_config.mxcfg" X$ Cluster = f$getsyi("CLUSTER_MEMBER") X$ X$ X$! --------\0808\1010\2020\400C X$! Set up the parameters; default if necessary X$! --------\0808\1010\2020\400C X$ Queues = P1 X$ if Queues .eqs. "" then $ \1909= "SYS$BATCH"`09! Batch queue for startup X$ Delay = P2 X$ if Delay .eqs. "" then $ \1909= "00:00:15"`09! Interval between tries X$ MaxTries = P3 X$ if MaxTries .eqs. "" then $ \1909= 240`09`09! Number of retries; an hour max Vimum X$ Username = P3 X$ if Username .eqs. "" then $ \1909= "SYSTEM"`09!\2F0Afor MX startup X$ X$ X$! --------\0808\1010\2020\400C X$! Check privileges required X$! --------\0808\1010\2020\400C X$ if .not. f$priviledge("SYSPRV,CMKRNL,SYSLCK") X$ then X$ em "-E-INSUFPRV, Insufficient privilege" X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Make sure we have access to the MCP program so we can shut down everything V X$! --------\0808\1010\2020\400C X$ if f$search("mx_exe:mcp.exe") .eqs. "" X$ then X$ em "-E-NOMCP, MX control program MCP not found" X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Check that we have a valid list of batch queue if CLUSTER operation has bee Vn X$! specified. The batch queues must execute on the nodes on which MX runs. X$! --------\0808\1010\2020\400C X$ if Cluster .eqs. "" X$ then X$ if P1 .eqs. "" X$ then X$ em "-E-CLUBATQUE, Batcxh execution queue list needed; member of clust Ver" X$ exit X$ endif X$ endif X$ X$! --------\0808\1010\2020\400C X$! Check the list of queues supplied to ensure that they are BATCH\2E08and X$! that they are EXECUTION queues rather than GENERIC\1B08- we want the X$! startup to execute on known nodes. X$! --------\0808\1010\2020\400C X$ n=0 X$QueueCheck: X$ q = f$element(n, ",", Queues) X$ if q .nes. "," X$ then X$ x = f$getqui("CANCEL_OPERATION") X$ x = f$getqui("DISPLAY_QUEUE", "QUEUE_NAME", q) X$ if x .eqs. "" X$ then X$ em "-E-QUENONEXI, ", q, " : not a valid queue name" X$ exit X$ endif X$ if .not. f$getqui("DISPLAY_QUEUE", "QUEUE_BATCH", q) X$ then X$ em "-E-QUEUNOTBAT, ", q, " : not a batch queue" X$ exit X$ endif X$ if f$getqui("DISPLAY_QUEUE", "QUEUE_GENERIC", q) X$ then X$ em "-E-QUENOTEXE, ", q, " : must be execution queue not generic" X$ exit X$ endif X$ n=n+1 X$ goto QueueCheck X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Shut all the MX processes having the queue file open X$! --------\0808\1010\2020\400C X$ if Cluster X$ then X$ em "-I-SHUTDOWN, Shutting down MX cluster wide X$ mcp shutdown /cluster X$ else X$ em "-I-SHUTDOWN, Shutting down MX" X$ mcp shutdown X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Loop until we can OPEN the SYSTEM_QUEUE file; when \2806, nobody X$! else is using it, which means all the agents have shutdown. X$! However, to avoid MX being shutdown forever due to any system problems X$! elsewhere, limit the amount of time we retry (see P2 and p3). X$! X$! NOTE: under the current version of MX (3.3 and below), the queue file X$! will be open if any user is in the process of sending an MX mail message. V X$! This is because MX opens the file before the user is asked to supply the X$! message and keeps it open throughout the entry of that \3707. X$! X$! --------\0808\1010\2020\400C X$ Tries = 0 X$ wait_loop: X$ if Tries .gt. Max\0E06then $ goto Wait_Too_Long X$ Tries =\0807+ 1 X$ wait 'Delay' X$ open/read/error=wait_loop tmp mx_flq_dir:system_queue.flq_ctl X$ close tmp X$ X$! --------\0808\1010\2020\400C X$! Now we can do the compress;\0A09 down to a temporary file first and X$! then rename it to the master one. This prevents unexpected aborts from X$! leaving partial queue files. X$! X$! NOTE: The FDL file used here should be tailored to the site by modifying X$! the ALLOCATION and EXTENSION values in AREA 0 and AREA 1. To avoid X$! excessive extent pointers building up, it is suggested that the X$! ALLOCATION quantity should be close to the maximum size reached by X$! the queue file. Excessive extent pointers will cause a serious X$! performance hit. X$! --------\0808\1010\2020\400C X$ em "-I-COMPRESS, Compressing MX queue file" X$ convert/fdl=mx_flq_dir:system_queue \1817.flq_ctl - X mx_flq_dir:system_queue.flq_ctl_tmp; X$ if .not. $status X$ then X$ em "-I-STATUS, ", $status, " : ", f$message(\1A07) X$ em "-I-COMPFAIL, Unable to complete the queue compress - aborted" X$ delete mx_flq_dir:system_queue.flq_ctl_tmp;* X$ goto CompressNotDone X$ endif X$ X$! --------\0808\1010\2020\400C X$! Rename the newly built queue file to the master\1306and purge off old X$! copies X$! --------\0808\1010\2020\400C X$ em "-I-UPDATE, Updating queue file" X$ rename mx_flq_dir:system_queue.flq_ctl_tmp \0D08 X$ purge mx_flq_dir:system_queue.flq_ctl X$ rename mx_flq_dir:system_queue.flq_ctl; ;1 X$ X$ X$! --------\0808\1010\2020\400C X$! Compress aborted, finished or timed out; restart MX X$! X$! The restart is complicated by two things. First, if the system is in a X$! cluster and the\1009mode is selected, it is necessary to restart MX X$! on all the nodes in the cluster that originally ran it. We do this by using V X$! either SYSMAN or a set of node specific BATCH queues. Second, this X$! procedure may be run by any user with the right privileges; however, MX mus Vt X$! be restarted under it's original username (see p4). If the requires\2309 X$! is the same as our current one, we use SYSMAN otherwise we submit the X$! startup procedure to each of the specified batch queues (see P1). X$! X$! --------\0808\1010\2020\400C X$CompressNotDone: X$Wait_Too_Long: X$RestartMx: X$ if f$edit(f$getjpi("","USERNAME"),"UPCASE,TRIM") .eqs. Username X$ then X$ X$ ! Issue the restart command on each node in the cluster using the sysman X$ ! command. Note that the MX_STARTUP_INFO file should specifically name X$ ! those nodes on which MX components should run because this procedure has V X$ ! no knowledge of which nodes in the cluster run MX and\2607don't. X$ em "-I-RESTART, Restarting MX on all nodes via SYSMAN" X$ mc sysman set env/cluster X set timeout 00:00:30`09`09&0209! In case of hangs! X do @sys$startup:mx_\0B07`09`09`09! Restart cluster-wide X$ else X$ X$ ! Start up the restart job by submitting it to a batch queue appropriate to V X$ ! the node (see p1) X$ n=0 X$ QueueSubmit: X$ q = f$element(n, ",", Queues) X$ if q .nes. "," X$ then X$ x = f$getqui("CANCEL_OPERATION") X$ node = f$getqui("DISPLAY_QUEUE", "SCSNODE_NAME", q, "ALL_JOBS") X$ em "-I-RESTART, Restarting MX on node ", node, " via job on batch que Vue ", q X$ submit/user='Username'/nolog/queue='q' sys$startup:mx_\0B07 X$ n=n+1 X$ goto QueueSubmit X$ endif X$ endif $ call unpack MX_SHRINK.COM;10 572365207 "" 18 1 2 $! $ create 'f' XMX QUEUE FILE COMPRESSION X X5-NOV-1993 X XAndy Harper`09`09&0209&0309Versiuon 1.2 X X--------\0808\1010\2020\400F X X XINTRODUCTION X XThe queue file used by the MX mail system (MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL) Xwill, on a heavily used system, grow unbounded. This causes performance Xproblems due to excessive use of file extent pointers (hence lots of disk Xseeks). X XThe DCL procedure described here is an attempt to reduce the performance Xproblems by periodically compressing the queue file. X X X XOPERATION X XThe queue file cannot be compressed while MX is running (strictly,\1F07any Xuser has the file open) so MX needs to be shut down. The procedure will first V Xshut MX down, then compress the queue file, and finally restart MX. Depending V Xon the size of the queue file, MX may be unavailable for some considerable Xtime. It is recommended that this procedure be run at the quietest moment Xovernight. X XThe procedure attempts to be fail safe. When the shutdown signal has been sen Vt Xto MX, it loops trying to open the queue file. Once it can, no-one is accessin Vg Xit and this indicates that MX has completely shut down (after finishing Xwhatever it was doing in a tidy manner). However, should the procedure be Xunable to open the queue file after a number of retries, it will give up and Xrestart MX. X XMX is also restarted if the procedure is aborted or\1F08compression attempt Xfails in some way. X XOnce the queue is accessible to the procedure, the compression operation Xbegins. A new file is created using the currently valid records of the existin Vg Xqueue. This new file is given a different name (and later renamed) to avoid a V Xpartial queue file being inadvertently put in place as the master. X XThe FDL file supplied with MX is used to create the queue\3206and it is Xstrongly recommended that the ALLOCATION and EXTENSION parameters of this file V Xbe tailored to the site; the exact values will depend on how much the queue Xfile grows during a typical day. It is recommended that the ALLOCATION Xparameters be set to regenerate the file close to its maximum size. X X X X XRUNNING THE PROCEDURE X XThe procedure is run as follows: X X @MX_SHRINK `5Bbatch-queues`5D `5Bdelay`5D `5Bretry`5D `5Busername`5D X Xwhere: X X `5Bbatch-queues`5D X`09`09Specifies a comma separated list of batch queues into which X`09`09the MX restart job is submitted. Each queue should execute on X`09`09one of the nodes in the cluster and there should be one queue X`09`09for each node on which MX runs. X X`09`09If the node is stand-alone, then this paramter is OPTIONAL X`09`09and defaults to SYS$BATCH. In a cluster, a suitable list X`09`09of batch queues is required. These must be \2806execution X`09`09queues. X X`09`09Note this parameter is NOT used if the procedure is submitted X`09`09under the username given by the `5B\1708`5D parameter. Instead, X`09`09the procedure is submitted using SYSMAN. X X `5Bdelay`5D`09is the delta time to wait between retries `5Bdefault: 15 seco Vnds`5D X X `5Bretry`5D`09is the number of times to retry`09if the queue file cannot be V X`09`09opened (default: 240) X X `5Busername`5D`09is the name under which to restart MX `5Bdefault: SYSTEM V`5D X X XYou should arrange to run this procedure on a regular basis every night by som Ve Xexternal means. X X X XACKNOWLEDGEMENTS X XI'm grateful to Hunter Goatley, who provided the basic draft of this procedure V, Xfor permission to make his code freely available in my modified form. X`0C XCHANGES FROM VERSION 1.1 X X1.`09SYSLCK privilege is required for an MX shutdown; this is now checked X`09before any actions are taken. X X2.`09The procedure now tried to deal with MX running in a cluster and using X`09a shared MX queue file by: X `09`09(a)`09Issuing a SHUTDOWN /CLUSTER command via MCP X`09`09(b)`09Restarting MX by submitting the MX_STARTUP job either X`09`09`09via SYSMAN or into a list of node specific batch queues X`09`09`09supplied as parameters. X X`09As a consequence, the parameter list of\1606rocedure has changed. X X`09See the documentation and source code for full details of how this X`09works. X X3.`09Fixed a typo in the documentation that transposed the parameters X`09`5Bretry`5D and `5Bdelay`5D. X X X X $ call unpack MX_SHRINK.DOC;3 54978660 "" 8 2 2 $ v=f$verify(v) $ exit ================================================================================ Archive-Date: Fri, 05 Nov 1993 13:22:04 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 14:17:34 EST From: system@aspentec.com Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <0097516E.F8A7ADA0.5587@aspentec.com> Subject: PROBLEM WITH THE SYSTEM_QUEUE FILE. We are running one copy of MX V3.3 on two AXP's and on one VAX. We have a SYSTEM_QUEUE.FLQ_CTL file that needs to have a PURGE and CONVERT on it at least two to three times a day. Because of the CONVERT we have to shutdown MX and have everyone exit e-mail to get a lock on the SYSTEM_QUEUE file. MX Router Router agent !ALPHA MX Local Local delivery agent MX Local Local delivery agent MX Local Local delivery agent MX MLF Mailing list/file server MX MLF Mailing list/file server MX MLF Mailing list/file server MX SMTP SMTP delivery agent MX SMTP SMTP delivery agent MX SMTP SMTP delivery agent MX uucp Intfc UUCP interface agent !VAX MX Site Agent Site-specific delivery agent MX Site Agent Site-specific delivery agent MX Site Agent Site-specific delivery agent SMTP Server SMTP server (over TCP/IP) SMTP Server SMTP server (over TCP/IP) SMTP Server SMTP server (over TCP/IP) These are the agents we are running, aside from the ROUTER running on one ALPHA and UUCP running on the VAX all nodes are running the same thing. ARE we missing something, is MX designed to run this way, are there any sites out there running like this? Please reply to me directly. Thanks. Jim MX%"system@aspentec.com" ================================================================================ Archive-Date: Fri, 05 Nov 1993 13:50:23 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: problem receiving mail on AXP/VMS system. Message-ID: <1993Nov5.165259.8525@dct.ac.uk> From: Reply-To: MX-List@WKUVX1.BITNET Date: 5 Nov 93 16:52:59 GMT To: MX-List@WKUVX1.BITNET In article <17215281@MVB.SAIC.COM>, figon@OBMARA.CNRS-MRS.FR writes: Hunter Goatley has pointed at the need to give the fully qualified domain name when installing UCX but what about this line: > X-MX-Warning: Warning -- Invalid "From" header. > From: Wed, 03 Nov 1993 22:14:57 EST -- Alan Greig Janet: A.Greig@uk.ac.dct Dundee Institute of Technology Internet: A.Greig@dct.ac.uk Tel: (0382) 308810 Int +44 382 308810 -- HELP -- ================================================================================ Archive-Date: Fri, 05 Nov 1993 14:30:06 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: PROBLEM WITH THE SYSTEM_QUEUE FILE. Message-ID: <1993Nov5.125242.286@buckie.hsc.colorado.edu> From: Date: 5 Nov 93 12:52:41 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17225010@MVB.SAIC.COM>, system@aspentec.com writes: >We are running one copy of MX V3.3 on two AXP's and on one VAX. >We have a SYSTEM_QUEUE.FLQ_CTL file that needs to have a PURGE and >CONVERT on it at least two to three times a day. [...] When was the last time you tuned the SYSTEM_QUEUE.FLQ_CTL file, so it has proper file extend sizes, bucket sizes, and the like? After you've done the initial tuning of the .FDL file, you can use Andy Harper's recently posted routine to automate having your file Converted nightly (note the other posts regarding shutting down MX cluster-wide and restarting MX cluster-wide, which you'll have to add to Andy's code until he releases an update). I don't remember the last time I shutdown MX and did a CONVERT of the queue file -- after I tuned it one time, things are moving along pretty well, and QUEUE SHOW/ALL commands complete pretty quickly, too (prior to tuning the file, a QUEUE SHOW/ALL command was taking 3-4 minutes to show a screenful of data -- now it takes 10 seconds to show a few screenfuls.) >Please reply to me directly. Posted and replied via Email. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Fri, 05 Nov 1993 15:47:33 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 15:40:53 CDT From: "John Clement-Rice U. (713)527-8101x2037" Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET CC: clement@physics.rice.edu Message-ID: <0097517A.9C860D80.17937@physics.rice.edu> Subject: Missing Date: field ? We are having a problem with mail that is sent by Eudora on a Mac, through MX. Essentialy Eudora is leaving off the Date: field. When I get a message from Eudora it does not have the field, while any message I send using MX or virtually any other mailer seems to have a Date field. 1. Does anyone know how to coerce Eudora into putting on the date??? or .... 2. Is it possible to have MX add the date after the fact ??? There are some mailers out there that can not handle this!!! MX does not seem to care, but I know of 1 site where the mail is bounced for lack of a date John Clement Rice University //\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ //John M. Clement Hepnet: UTSPAN::RIPHYS::CLEMENT //Bonner Lab Bitnet: CLEMENT@RICEVM1 //Rice University Internet: CLEMENT@PHYSICS.RICE.EDU //PO Box 1892 BellTel: (713)527-8181 x2037 //Houston, Tx. 77251 Fax: (713)285-5215 //\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ ================================================================================ Archive-Date: Fri, 05 Nov 1993 16:02:55 CST Sender: list-mgr@WKUVX1.BITNET Subject: HELP - MX queue gone mad (FIN)!!! Message-ID: <1993Nov5.113652.1@woods.ulowell.edu> From: desroches@woods.ulowell.edu Reply-To: MX-List@WKUVX1.BITNET Date: 5 Nov 93 11:36:52 -0500 To: MX-List@WKUVX1.BITNET HELP!!! We had a brief power failure Sunday night. For various reasons MX wasn't restarted until Tuesday morning. Since then the QUEUE has been full of messages whose status is FIN. There are approximately 8000 of them and they won't go away!!!!!!!!! I tried doing a QUEUE PURGE/LOG but this took about 1 minute to purge a single entry. I also did a RECLAIM but that had no effect. We are running v3.1. I'd appreciate any help anyone can give. Also, is there an official person or site to whom I could address such a question? I know that used to be Matt Madison and now I think it may be Hunter Goately but I'm not sure if they will accept direct questions. Sorry if these are dumb question but I haven't been near MX in years. ================================================================================ Archive-Date: Fri, 05 Nov 1993 17:05:00 CST Sender: list-mgr@WKUVX1.BITNET Date: Fri, 05 Nov 1993 16:44:43 CST From: "George D. Greenwade" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU CC: decroches@woods.ulowell.edu Message-ID: <00975183.87557780.2735@SHSU.edu> Subject: RE: HELP - MX queue gone mad (FIN)!!! On 5 Nov 93 11:36:52 -0500, desroches@woods.ulowell.edu posted: > HELP!!! We had a brief power failure Sunday night. For various reasons MX > wasn't restarted until Tuesday morning. > > Since then the QUEUE has been full of messages whose status is FIN. There > are approximately 8000 of them and they won't go away!!!!!!!!! I tried > doing a QUEUE PURGE/LOG but this took about 1 minute to purge a single > entry. I also did a RECLAIM but that had no effect. > > We are running v3.1. I'd appreciate any help anyone can give. Your best bet is to shutdown MX as your SYSTEM_QUEUE.FLQ_CTL file is probably large and is getting locked by a variety of things. Also, in 3.1 the FLQ_DIR was a single entity and this means all kinds of headaches. Once shutdown, do your MCP QUE PURGE -- it'l take some time, but you should have a pretty unique lock on everything. After that, update to 3.3 as the FLQ_DIR (MX_FLQ_DIR in 3.3) is a lot better structured to handle things along these lines. > Also, is there an official person or site to whom I could address such a > question? I know that used to be Matt Madison and now I think it may be > Hunter Goately but I'm not sure if they will accept direct questions. MX-List is about as good as you can do. Regards and good luck, George %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% George D. Greenwade, Ph.D. Bitnet: BED_GDG@SHSU Department of Economics and Business Analysis THEnet: SHSU::BED_GDG College of Business Administration Voice: (409) 294-1266 P. O. Box 2118 FAX: (409) 294-3612 Sam Houston State University Internet: bed_gdg@SHSU.edu Huntsville, TX 77341 bed_gdg%SHSU.decnet@relay.the.net %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ================================================================================ Archive-Date: Fri, 05 Nov 1993 18:56:06 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: Missing Date: field ? Message-ID: <1993Nov5.151625.291@buckie.hsc.colorado.edu> From: Date: 5 Nov 93 15:16:25 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <17229247@MVB.SAIC.COM>, "John Clement-Rice U. (713)527-8101x2037" writes: >We are having a problem with mail that is sent by Eudora on a Mac, >through MX. Essentialy Eudora is leaving off the Date: field. >When I get a message from Eudora it does not have the field, while any message >I send using MX or virtually any other mailer seems to have a Date field. > >1. Does anyone know how to coerce Eudora into putting on the date??? > or .... >2. Is it possible to have MX add the date after the fact ??? > >There are some mailers out there that can not handle this!!! >MX does not seem to care, but I know of 1 site where the mail is bounced >for lack of a date You can write up a simple MX SITE agent, in DCL even, which will add the "Date:" after the fact -- should be pretty trivial, but a bit slow. You'd have to force all mail leaving the Macs to go through your node running MX, of course. Have you contacted the maintainer of Eudora for the Mac to see if the bug has already been fixed? -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Fri, 05 Nov 1993 19:32:26 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: HELP - MX queue gone mad (FIN)!!! Message-ID: <1993Nov5.150144.290@buckie.hsc.colorado.edu> From: Date: 5 Nov 93 15:01:43 MDT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <1993Nov5.113652.1@woods.ulowell.edu>, desroches@woods.ulowell.edu writes: >HELP!!! We had a brief power failure Sunday night. For various reasons MX >wasn't restarted until Tuesday morning. > >Since then the QUEUE has been full of messages whose status is FIN. There >are approximately 8000 of them and they won't go away!!!!!!!!! I tried >doing a QUEUE PURGE/LOG but this took about 1 minute to purge a single >entry. I also did a RECLAIM but that had no effect. I'd recommend you shutdown all MX agents and tune your queue file (MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL). Andy Harper just posted a new tuning program that should work quite well (I haven't tried it yet), and there's also one in MX_CONTRIB, I believe (if the person that installed MX on your system included the Contributed software, that is.) >We are running v3.1. I'd appreciate any help anyone can give. When convenient, you might want to upgrade to V3.3 -- it is the current version, and includes some performance gains over V3.1 (the performance gains were actually added in V3.2). >Also, is there an official person or site to whom I could address such a >question? I know that used to be Matt Madison and now I think it may be >Hunter Goately but I'm not sure if they will accept direct questions. The "best" thing, and I'm sure Hunter would agree, is to throw them out on the list MX-LIST@WKUVX1.BITNET or post them in news to vmsnet.mail.mx -- they are bi-directionally gatewayed to each other. If you post to the MX list, someone, somewhere, will be awake and available to answer your question; if you send mail to Hunter, he could be out helping his wife delivery babies or something! ;-) -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Sat, 06 Nov 1993 09:15:19 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: HELP - MX queue gone mad (FIN)!!! Message-ID: <1993Nov6.125316.10038@aragorn.unibe.ch> From: Date: Sat, 6 Nov 1993 12:53:16 GMT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <1993Nov5.113652.1@woods.ulowell.edu>, desroches@woods.ulowell.edu writes: >HELP!!! We had a brief power failure Sunday night. For various reasons MX >wasn't restarted until Tuesday morning. > >Since then the QUEUE has been full of messages whose status is FIN. There >are approximately 8000 of them and they won't go away!!!!!!!!! I tried >doing a QUEUE PURGE/LOG but this took about 1 minute to purge a single >entry. I also did a RECLAIM but that had no effect. I had this situation once. You have to: a) shut down all MX processes b) do a QUEUE CANCEL xxxx for all entries in the queue c) do a QUEUE PURGE d) do a RECLAIM e) restart MX at least, this procedure solved the problem here. >Also, is there an official person or site to whom I could address such a >question? I know that used to be Matt Madison and now I think it may be >Hunter Goately but I'm not sure if they will accept direct questions. This group is the best place to put questions. Martin ******************************************************************************* Martin Egger, Ph.D., Computing Services - Head of System/User Support Group University of Bern, Gesellschaftsstrasse 6, CH-3012 Bern, Switzerland Phone: ++41 (0)31 631 38 45, Fax: ++41 (0)31 631 38 65, Telex: 912643 pibe ch RFC: egger@id.unibe.ch, X.400: S=egger;OU=id;O=unibe;P=switch;A=arcom;C=ch; HEPNET/SPAN: 20579::49202::egger, DECnet (Switzerland): 49202::egger ******************************************************************************* ================================================================================ Archive-Date: Sat, 06 Nov 1993 11:26:23 CST Sender: list-mgr@WKUVX1.BITNET Subject: Re: DECnet addresses in "To" line. Message-ID: <2b66do$kd2@lohengrin.umkc.edu> From: Date: 2 Nov 1993 17:46:32 GMT Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET In article <2amras$f3q@lohengrin.umkc.edu>, mckeever@CSTP.UMKC.EDU (Brian McKeever) writes: >I've got addresses that look like this: > > To: "cstp::mckeever"@PRIMUS.CSTP.UMKC.EDU > >I would really like to do away with the DECnet portion of the address >and replace it with just the username. "cstp::" is the cluster alias >name not the nodes name (PRIMUS::). > >I'm _sure_ this has probably been covered before, but I took a couple >of hours and searched through the MX archives but couldn't find >anything that looked appropriate. I don't think I've ever seen anyone answer their own post before! Anyway... It doesn't look like this problem, in exactly this form, has been addressed before. Dan Wing (dwing@uh01.Colorado.EDU) sent me a courteous reply to my original post indicating that my MAIL$SYSTEM_FLAGS might be set incorrectly. I had already checked that, but I did double check. It is set to "7". He was also kind enough to look at my routes and paths. It all looked fine to him too. I then sent him _exactly_ the commands I use to create the above type of address, and his configuration does it too. Here's how to reproduce the problem at will (a full transcript is included at the end of this posting): 1) Mail yourself a message without using the MX% prefix. This just gets delivered. No MX involved at all. 2) Read the message you just sent to yourself and REPLY to it CC'ing yourself at the address MX%. After a few moments, you should have two messages. The first message should be a the standard VMS Mail message and the second one from MX. The MX message will have the bogus "To:" field in it. The sad thing is that this bad address seems to sometimes get sent to a remote host resulting in a non-replyable message (I haven't been able to reproduce this behavior yet so I don't know under what circumstances it happens). I turned on MX debugging on the router, and it doesn't even look like MX is looking at the address it writes in the "To:" line in the message header: 1-NOV-1993 12:35:35.78 %PROCESS, Processing entry number 5500 1-NOV-1993 12:35:35.86 %PROCESS, Status from READ_INFO was 00000001 1-NOV-1993 12:35:35.86 %PROCESS, Message originated in VMS Mail. 1-NOV-1993 12:35:35.86 %PROCESS, will run domain expander on envelope addresses. 1-NOV-1993 12:35:35.86 %PROCESS, Processing address: 1-NOV-1993 12:35:35.87 %PROCESS, ... address now reads: 1-NOV-1993 12:35:35.87 %PROCESS, will run domain expander on message headers. 1-NOV-1993 12:35:36.46 %PROCESS, Finished VMSmail-origin preprocessing. 1-NOV-1993 12:35:36.46 %PROCESS, Recipient #0: 1-NOV-1993 12:35:36.46 %REWRITE, No rewrite rules matched 1-NOV-1993 12:35:36.47 %FINDPATH, Site-spec expand on PRIMUS.CSTP.UMKC.EDU err=00000000 1-NOV-1993 12:35:36.47 %FINDPATH, domain name PRIMUS.CSTP.UMKC.EDU matched path pattern PRIMUS.CSTP.UMKC.EDU 1-NOV-1993 12:35:36.47 %PROCESS, Rewrote as - next hop PRIMUS.CSTP.UMKC.EDU, path 1 1-NOV-1993 12:35:36.47 %FINDALIAS, no alias found for mckeever 1-NOV-1993 12:35:36.47 %PROCESS, no alias found for mckeever 1-NOV-1993 12:35:36.47 %PROCESS, this is just a local delivery 1-NOV-1993 12:35:36.54 %PROCESS, Adding to LOCAL path: mckeever. 1-NOV-1993 12:35:36.77 %PROCESS, Path LOCAL gets 1 rcpts, entry number 5501 The following rewrite rule has no affect: Address-rewriting rules: Rewrite "" => "<{user}@{host}>" Nor did this one: Address-rewriting rules: Rewrite "<""cstp::{user}""@{host}>" => "<{user}@{host}>" Interesting enough, when you mix a local and a MX address with SEND instead of REPLY, no bogus addresses appear. So... Is this a bug? Feature? OpenVMS/AXP V1.5, MX V3.3. Thanks, -=Brian +---- Full REPLY transcript ------------------------------------------+ PRIMUS> MAIL MAIL> SEND/NOEDIT To: MCKEEVER CC: Subj: TEST... Enter your message below. Press CTRL/Z when complete, or CTRL/C to quit: TEST... -=Brian New mail on node PRIMUS from CSTP::MCKEEVER "Brian McKeever" (11:31:32) MAIL> read/new #1 2-NOV-1993 11:31:32.37 NEWMAIL From: CSTP::MCKEEVER "Brian McKeever" To: MCKEEVER CC: Subj: TEST... TEST... -=Brian MAIL> REPLY/NOEDIT To: CSTP::MCKEEVER CC: MX%MCKEEVER Subj: RE: TEST... Enter your message below. Press CTRL/Z when complete, or CTRL/C to quit: TEST... (REPLY) -=Brian New mail on node PRIMUS from CSTP::MCKEEVER "Brian McKeever" (11:31:56) Press RETURN to continue reading your mail New mail on node PRIMUS from MX%"mckeever@PRIMUS.CSTP.UMKC.EDU" (11:31:58) MAIL> read/new #1 2-NOV-1993 11:31:55.82 NEWMAIL From: CSTP::MCKEEVER "Brian McKeever" To: CSTP::MCKEEVER CC: MX%MCKEEVER Subj: RE: TEST... TEST... (REPLY) -=Brian MAIL> n #2 2-NOV-1993 11:31:57.99 NEWMAIL From: MX%"mckeever@PRIMUS.CSTP.UMKC.EDU" To: MCKEEVER CC: Subj: RE: TEST... Return-Path: Received: by PRIMUS.CSTP.UMKC.EDU (MX V3.3 AXP) id 5743; Tue, 02 Nov 1993 11:31:56 CDT Date: Tue, 02 Nov 1993 11:31:56 CDT From: Brian McKeever To: "cstp::mckeever"@PRIMUS.CSTP.UMKC.EDU CC: mckeever@PRIMUS.CSTP.UMKC.EDU Message-ID: <00974EFC.55CA2AB4.5743@PRIMUS.CSTP.UMKC.EDU> Subject: RE: TEST... TEST... (REPLY) -=Brian MAIL> EXIT +---- Full REPLY transcript ------------------------------------------+ ================================================================================ Archive-Date: Sat, 06 Nov 1993 13:57:05 CST Sender: list-mgr@WKUVX1.BITNET Date: Sat, 06 Nov 1993 12:54:08 MST From: "Michael L. Hitch" Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU Message-ID: <0097522C.7B193FC0.1641@msu.oscs.montana.edu> Subject: Re: DECnet addresses in "To" line. writes: > In article <2amras$f3q@lohengrin.umkc.edu>, mckeever@CSTP.UMKC.EDU (Brian > McKeever) writes: > >I've got addresses that look like this: > > > > To: "cstp::mckeever"@PRIMUS.CSTP.UMKC.EDU > > ... > it. The sad thing is that this bad address seems to sometimes get sent > to a remote host resulting in a non-replyable message (I haven't been > able to reproduce this behavior yet so I don't know under what > circumstances it happens). I think the " characters can confuse some mailers. We have an HP workstation in our library that is used as a mail gateway to the library's HP system. I think the HP tries to escape the " characters, which mangles the address and causes a lot of confusion. Michael --- Michael L. Hitch osymh@msu.oscs.montana.edu Computer Consultant OSYMH@MTSUNIX1.BITNET Office of Systems and Computing Services Montana State University Bozeman, MT USA ================================================================================ Archive-Date: Sat, 06 Nov 1993 18:06:18 CST Sender: list-mgr@WKUVX1.BITNET Date: Sun, 07 Nov 1993 00:02:09 GMT From: Jamie Jones Reply-To: MX-List@WKUVX1.BITNET To: mx-list@WKUVX1.BITNET Message-ID: <00975289.CDA8D4A0.40@hicom.lut.ac.uk> Subject: Re: DECnet addresses in "To" line. => From: [ ... ] => 1) Mail yourself a message without using the MX% prefix. This => just gets delivered. No MX involved at all. => 2) Read the message you just sent to yourself and REPLY to it => CC'ing yourself at the address MX%. After a few => moments, you should have two messages. => => The first message should be a the standard VMS Mail message and the => second one from MX. The MX message will have the bogus "To:" field in => it. The sad thing is that this bad address seems to sometimes get sent => to a remote host resulting in a non-replyable message (I haven't been => able to reproduce this behavior yet so I don't know under what => circumstances it happens). Coincidently, I've noticed this on our non-clustered machine recently when replying to internal mail with mail default transport set to MX% (I know, serves me right for playing with an undocumented command... :-) ) Basically, its like saying : MAIL> send To: MX%HICOM::A_USER .. MX (V3.3 here, too) puts the header "To:" line as : To: "hicom::a_user"@hicom.lut.ac.uk Now, if this is CC'd externally, such as : MAIL> send *** or this could be a MAIL> reply *** To: MX%HICOM::A_USER *** To: HICOM::A_USER *** CC: MX%"someone@fishing.net" The header (which would also be sent to the remote site) would have : To: "hicom::a_user"@hicom.lut.ac.uk CC: someone@fishing.net Is this what causes problems ? It would be nice if this could be 'fixed' purely for asthetic purposes :) Jamie. -- Jamie Jones, HICOM Operations Group ================================================================================ Archive-Date: Sat, 06 Nov 1993 20:37:02 CST Sender: list-mgr@WKUVX1.BITNET Date: Sat, 06 Nov 1993 20:36:18 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.BITNET To: MX-List@WKUVX1.BITNET Message-ID: <0097526D.0BB1833B.17254@WKUVX1.BITNET> Subject: Re: DECnet addresses in "To" line. Jamie Jones writes: > >To: "hicom::a_user"@hicom.lut.ac.uk >CC: someone@fishing.net > >Is this what causes problems ? >It would be nice if this could be 'fixed' purely for asthetic purposes :) > Just in case anyone wonders, the address shown above *is* a valid address. The problem with trying to "fix" it is that the node information could be necessary to get replies back to the user.... Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKUVX1.BITNET (or goathunter%wkuvx1.bitnet@ULKYVM.LOUISVILLE.EDU) ================================================================================ Archive-Date: Mon, 08 Nov 1993 02:02:23 CST Sender: list-mgr@WKUVX1.BITNET Date: Mon, 08 Nov 1993 09:02:55 EST From: figon@OBMARA.CNRS-MRS.FR Reply-To: MX-List@WKUVX1.BITNET To: MX-List%WKUVX1.bitnet@ULKYVM.LOUISVILLE.EDU CC: figon@OBMARA.CNRS-MRS.FR Message-ID: <0097539E.831C9226.23618@OBMARA.CNRS-MRS.FR> Subject: Re: problem receiving mail on AXP/VMS system. THANK YOU, ALPHAB IS NOT DEFINED AS A LOCAL DELIVERY, IT IS OK NOW. P.F ================================================================================ Archive-Date: Mon, 08 Nov 1993 17:17:30 CST Sender: list-mgr@WKU.EDU Date: Mon, 08 Nov 1993 16:58:03 EST From: "Andy, Systems Manager" Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU CC: udaa055@BAY.CC.KCL.AC.UK Message-ID: <009753E0.E3C187C0.11191@bay.cc.kcl.ac.uk> Subject: MX queue not auto-purging I seem to have a situation whereby my MX queue file is not automatically clearing out the old FINISHED entries! As a result, my queue file grows quite large! I do not have the appropriate MX logical defined and, according to the documentation, it should default to clearing out every 15 minutes. Could anyone suggest what the problem might be? At the moment I'm having to do a manual QUEUE PURGE and it can take a long long time if you don't do it frequently! I'm running MX 3.3 on VMS 5.5-2 Thanks Andy Harper Kings College London ================================================================================ Archive-Date: Mon, 08 Nov 1993 17:41:19 CST Sender: list-mgr@WKU.EDU Date: Mon, 08 Nov 1993 15:33:55 +0100 From: Herve GILIBERT CRI universite Jean Monnet ST ETIENNE Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU Message-ID: <009753D5.22575FC0.17841@stroph.univ-st-etienne.fr> Subject: MX accounting. Last week in a question I asked about MX accounting I mentioned that a new version of accounting file was created every one or two days. Now the creation of this new version seems to appear when the name server (which is running on the same machine than MX) stops running. I never found why my nameserver was stopping every one or two days. Does somebody experiences such a situation ? MX 3.3 CMUIP 6.5-A with Bruce Orchard name server. Thanks in advance. -- Herve GILIBERT UUUU UUUU CRITeR UUUU UUUU Universite Jean Monnet St-Etienne UUUU UUUU 23, Rue du Dr. P. Michelon UUUU UUUU 42023 ST ETIENNE Cedex 2 UUUU UUUU Tel : 33 77 42 15 79 UUUU UUUU Fax : 33 77 42 15 75 UUUUUUUUUUU RFC822: gilibert@criter.univ-st-etienne.fr UUUUUUUUU ================================================================================ Archive-Date: Mon, 08 Nov 1993 22:20:32 CST Sender: list-mgr@WKU.EDU Date: Mon, 08 Nov 1993 21:58:42 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <0097540A.E3A3FA9B.17961@ALPHA.WKU.EDU> Subject: RE: MX queue not auto-purging As you can see, WKUVX1 is now (finally!) on the Internet! "Andy, Systems Manager" writes: > >I seem to have a situation whereby my MX queue file is not automatically >clearing out the old FINISHED entries! As a result, my queue file grows quite >large! I do not have the appropriate MX logical defined and, according to the >documentation, it should default to clearing out every 15 minutes. > >Could anyone suggest what the problem might be? At the moment I'm having to do >a manual QUEUE PURGE and it can take a long long time if you don't do it >frequently! > I had this same thing happen to me today. The problem here, and most likely yours, is that there are so many messages to be processed that the Router never gets finished with new messages, so it never gets around to the PURGE. Doing it by hand a few times is one way; running a second Router process may help. I've started looking seriously at these indexed file performance problems. No matter how I look at it, it looks complex. Still, I do intend to move away from the indexed file in, hopefully, the next version of MX. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Tue, 09 Nov 1993 08:01:34 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 09 Nov 1993 08:52:26 EST From: "Brian Tillman" Reply-To: MX-List@WKUVX1.WKU.EDU To: WKU.EDU!mx-list@esseye.si.com Message-ID: <00975466.36D2EFA0.12922@swdev.si.com> Subject: RE: MX queue not auto-purging Hunter Goatley, (goathunter@WKU.EDU) writes: >As you can see, WKUVX1 is now (finally!) on the Internet! Yes, and FILESERV sent my most recent request all at once. Nice! >"Andy, Systems Manager" writes: >> >>I seem to have a situation whereby my MX queue file is not automatically >>clearing out the old FINISHED entries! As a result, my queue file grows quite >>large! I do not have the appropriate MX logical defined and, according to the >>documentation, it should default to clearing out every 15 minutes. >> >>Could anyone suggest what the problem might be? At the moment I'm having to do >>a manual QUEUE PURGE and it can take a long long time if you don't do it >>frequently! >> >I had this same thing happen to me today. The problem here, and most >likely yours, is that there are so many messages to be processed that >the Router never gets finished with new messages, so it never gets >around to the PURGE. This very thing also happened to me yesterday. A _ton_ of messages came in and, even though I run two routers (one on each of two machines in a cluster), one router was doing nothing but purging finished entries. I had to wait about 40 minutes after a SHUTDOWN for one of the routers to finish what it was doing. Trying to use MCP QUEUE SHOW during this time was a total waste of time. -- Brian Tillman tillman_brian@si.com ================================================================================ Archive-Date: Tue, 09 Nov 1993 08:40:42 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 09 Nov 1993 08:37:45 CST From: "Hunter Goatley, WKU" Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975464.29C74295.18359@ALPHA.WKU.EDU> Subject: RE: MX queue not auto-purging "Brian Tillman" writes: > >Hunter Goatley, (goathunter@WKU.EDU) writes: >>As you can see, WKUVX1 is now (finally!) on the Internet! > >Yes, and FILESERV sent my most recent request all at once. Nice! > Good deal! >This very thing also happened to me yesterday. A _ton_ of messages came in and, >even though I run two routers (one on each of two machines in a cluster), one >router was doing nothing but purging finished entries. I had to wait about 40 >minutes after a SHUTDOWN for one of the routers to finish what it was doing. >Trying to use MCP QUEUE SHOW during this time was a total waste of time. Yeah. Pretty frustrating, huh? As I said, I hope to do something about it in the next release. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Tue, 09 Nov 1993 09:49:03 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 09 Nov 1993 15:21:26 EST From: "Andy, Systems Manager" Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU CC: udaa055@BAY.CC.KCL.AC.UK Message-ID: <0097549C.8F3BC780.13954@bay.cc.kcl.ac.uk> Subject: MX queue compression - MX_SHRINK V1.4 I've improved the MX_SHRINK utility, recently posted, by adding some code to optimize the size of the MX queue file. This is handled by using the optimize script of the EDIT/FDL command. The latest code is enclosed below. Please let me know if you find any problems with it. Thanks to Dan Wing for a number of useful improvements and for pointing out a number of typos in the code and docs. Regards, Andy Harper Kings College London $! ------------------ CUT HERE ----------------------- $ v='f$verify(f$trnlnm("SHARE_UNPACK_VERIFY"))' $! $! This archive created: $! Name : MX_SHRINK $! By : Andy Harper $! Date : 9-NOV-1993 15:05:33.46 $! Using: VMS_SHARE 8.4, (C) 1993 Andy Harper, Kings College London UK $! $! Credit is due to these people for their original ideas: $! James Gray, Michael Bednarek $! $! TO UNPACK THIS SHARE FILE, CONCATENATE ALL PARTS IN ORDER $! AND EXECUTE AS A COMMAND PROCEDURE ( @name ) $! $! THE FOLLOWING FILE(S) WILL BE CREATED AFTER UNPACKING: $! 1. MX_SHRINK.COM;17 $! 2. MX_SHRINK.DOC;6 $! $ set="set" $ set symbol/scope=(nolocal,noglobal) $ f=f$parse("SHARE_UNPACK_TEMP","SYS$SCRATCH:."+f$getjpi("","PID")) $ e="write sys$error ""%UNPACK"", " $ w="write sys$output ""%UNPACK"", " $ if .not. f$trnlnm("SHARE_UNPACK_LOG") then $ w = "!" $ if f$getsyi("CPU") .gt. 127 then $ goto start $ ve=f$getsyi("version") $ if ve-f$extract(0,1,ve) .ges. "4.4" then $ goto start $ e "-E-OLDVER, Must run at least VMS 4.4" $ v=f$verify(v) $ exit 44 $unpack: subroutine ! P1=filename, P2=checksum, P3=attributes,P4=size $ if f$parse(P1) .nes. "" then $ goto dirok $ dn=f$parse(P1,,,"DIRECTORY") $ w "-I-CREDIR, Creating directory ''dn'" $ create/dir 'dn' $ if $status then $ goto dirok $ e "-E-CREDIRFAIL, Unable to create ''dn' File skipped" $ delete 'f'* $ exit $dirok: $ x=f$search(P1) $ if x .eqs. "" then $ goto file_absent $ e "-W-EXISTS, File ''P1' exists. Skipped" $ delete 'f'* $ exit $file_absent: $ w "-I-UNPACK, Unpacking ", P5, " of ", P6, " - ", P1, " - ", P4, " Blocks" $ n=P1 $ if P3 .nes. "" then $ n=f $ if .not. f$verify() then $ define/user sys$output nl: $ EDIT/TPU/NOSEC/NODIS/COM=SYS$INPUT/NOJOURNAL 'f'/OUT='n' PROCEDURE GetHex(s,p)LOCAL x1,x2;x1:=INDEX(t,SUBSTR(s,p,1))-1;x2:=INDEX(t, SUBSTR(s,p+1,1))-1;RETURN 16*x1+x2;ENDPROCEDURE;PROCEDURE SkipPartsep LOCAL m; LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m);EXITIF INDEX(ERASE_LINE, "-+-+-+-+-+-+-+-+")=1;ENDLOOP;ENDPROCEDURE;PROCEDURE ProcessLine LOCAL c,s,l,b, n,p;c := ERASE_CHARACTER(1);s := ERASE_LINE;EDIT(s,"TRIM");IF c = "X" THEN SPLIT_LINE; ENDIF;MOVE_HORIZONTAL(-1);l := LENGTH(s);p := 1; LOOP EXITIF p > l;c := SUBSTR(s,p,1);p := p+1;CASE c FROM ' ' TO '`' ['\']: b:= GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT( SUBSTR(CURRENT_LINE, CURRENT_OFFSET-b+1,n));['&']: b:=GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT(ASCII(n)*b);['`']: COPY_TEXT(ASCII(GetHex(s,p))); p:=p+2;[INRANGE, OUTRANGE]: COPY_TEXT(c);ENDCASE;ENDLOOP;ENDPROCEDURE;PROCEDURE Decode LOCAL m; POSITION(BEGINNING_OF(b));LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m); IF INDEX(CURRENT_LINE,"+-+-+-+-+-+-+-+-")=1 THEN SkipPartSep;ELSE ProcessLine; MOVE_HORIZONTAL(1);ENDIF;ENDLOOP;ENDPROCEDURE;SET(FACILITY_NAME,"UNPACK");SET( SUCCESS,OFF);SET(INFORMATIONAL,OFF);t:="0123456789ABCDEF";f:=GET_INFO( COMMAND_LINE,"file_name");b:=CREATE_BUFFER(f,f);Decode;WRITE_FILE(b,GET_INFO( COMMAND_LINE,"output_file"));QUIT; $ if p3 .eqs. "" then $ goto dl $ open/write fdl &f $ write fdl "RECORD" $ write fdl P3 $ close fdl $ w "-I-CONVRFM, Converting record format to ", P3 $ convert/fdl=&f &f-1 &P1 $dl: delete 'f'* $ checksum 'P1' $ if checksum$checksum .nes. P2 then $ - e "-E-CHKSMFAIL, Checksum of ''P1' failed." $ exit $ endsubroutine $start: $! $ create 'f' X$! Purpose: X$!`09Automatically compress the MX QUEUE file X$! X$! Parameters: X$!`09P1`09List of queues for restart job. if a stand-alone node, this X$!`09`09is optional and defaults to SYS$BATCH. If a member of a cluster X$!`09`09a list of execution batch queues (one per node) is required X$!`09`09This parameter is ignored if P4 (see below) is the same as the X$!`09`09current user. SYSMAN is used to restart MX in this case. X$!`09P2`09`5Boptional`5D delay time between queue file access retries (default V: 15 seconds) X$!`09P3`09`5Boptional`5D number of retries (default: 240 ) X$!`09P4`09`5Boptional`5D username to run MX (default: SYSTEM) X$! X$! Privileges: X$!`09SYSPRV X$!`09CMKRNL X$!`09SYSLCK X$!`09SETPRV X$!`09 X$! Environment: X$!`09Requires MX logicals to be defined X$!`09Requires MX software to be present X$! X$! Revision History: X$!`091.0`09??????????`09Hunter Goatley`09Original version X$!`091.1`091-NOV-1993`09Andy Harper`09Added parameters, comments, error checki Vng X$!`091.2`095-NOV-1993`09Andy Harper`09Add cluster support; fix privs X$!`091.3`098-NOV-1993`09Andy Harper`09Minor typos; add queue purge X$!`091.4`098-NOV-1993`09Andy harper`09Add automatic optimization X$! X$ X$ On control_y then $ goto RestartMX X$ set noon X$ em = "write sys$output ""%"+f$parse(f$environment("PROCEDURE"),,,"NAME")+""" V," X$ mcp = "$mx_exe:mcp.exe/file=mx_dir:mx_config.mxcfg" X$ Cluster = f$getsyi("CLUSTER_MEMBER") X$ X$ MXqueue = "MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL" X$ MXfdl = "MX_FLQ_DIR:SYSTEM_QUEUE.FDL" X$ X$ X$! --------\0808\1010\2020\400C X$! Set up the parameters; default if necessary X$! --------\0808\1010\2020\400C X$ Queues = P1 X$ if Queues .eqs. "" then $ \1909= "SYS$BATCH"`09! Batch queue for startup X$ Delay = P2 X$ if Delay .eqs. "" then $ \1909= "00:00:15"`09! Interval between tries X$ MaxTries = P3 X$ if MaxTries .eqs. "" then $ \1909= 240`09`09! Number of retries; an hour max Vimum X$ Username = P3 X$ if Username .eqs. "" then $ \1909= "SYSTEM"`09!\2F0Afor MX startup X$ X$ X$! --------\0808\1010\2020\400C X$! Check privileges required X$! --------\0808\1010\2020\400C X$ if .not. f$priviledge("SYSPRV,CMKRNL,SYSLCK") X$ then X$ em "-E-INSUFPRV, Insufficient privilege" X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Make sure we have access to the MCP program so we can shut down everything V X$! --------\0808\1010\2020\400C X$ if f$search("mx_exe:mcp.exe") .eqs. "" X$ then X$ em "-E-NOMCP, MX control program MCP not found" X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Check that we have a valid list of batch queue if CLUSTER operation has bee Vn X$! specified. The batch queues must execute on the nodes on which MX runs. X$! --------\0808\1010\2020\400C X$ if Cluster .eqs. "" X$ then X$ if P1 .eqs. "" X$ then X$ em "-E-CLUBATQUE, Batch execution queue list needed; member of cluste Vr" X$ exit X$ endif X$ endif X$ X$! --------\0808\1010\2020\400C X$! Check the list of queues supplied to ensure that they are BATCH\2E08and X$! that they are EXECUTION queues rather than GENERIC\1B08- we want the X$! startup to execute on known nodes. X$! --------\0808\1010\2020\400C X$ n=0 X$QueueCheck: X$ q = f$element(n, ",", Queues) X$ if q .nes. "," X$ then X$ x = f$getqui("CANCEL_OPERATION") X$ x = f$getqui("DISPLAY_QUEUE", "QUEUE_NAME", q) X$ if x .eqs. "" X$ then X$ em "-E-QUENONEXI, ", q, " : not a valid queue name" X$ exit X$ endif X$ if .not. f$getqui("DISPLAY_QUEUE", "QUEUE_BATCH", q) X$ then X$ em "-E-QUEUNOTBAT, ", q, " : not a batch queue" X$ exit X$ endif X$ if f$getqui("DISPLAY_QUEUE", "QUEUE_GENERIC", q) X$ then X$ em "-E-QUENOTEXE, ", q, " : must be execution queue not generic" X$ exit X$ endif X$ n=n+1 X$ goto QueueCheck X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Do a QUEUE PURGE to force out as many dead entries as possible. Note that X$! this should normally get done automatically by the ROUTER but sometimes it V X$! doesn't so this is a safety precaution. Note that it may take a LONG time o Vn X$! some systems and you may want to comment it out!!! X$! --------\0808\1010\2020\400C X$ em "-I-PURGING, Purging old entries" X$ mc mx_exe:mcp queue purge X$ X$ X$! --------\0808\1010\2020\400C X$! Shut all the MX processes having the queue file open X$! --------\0808\1010\2020\400C X$ if Cluster X$ then X$ em "-I-SHUTDOWN, Shutting down MX cluster wide X$ mcp shutdown /cluster X$ else X$ em "-I-SHUTDOWN, Shutting down MX" X$ mcp shutdown X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Loop until we can OPEN the SYSTEM_QUEUE file; when \2806, nobody X$! else is using it, which means all the agents have shutdown. X$! However, to avoid MX being shutdown forever due to any system problems X$! elsewhere, limit the amount of time we retry (see P2 and p3). X$! X$! NOTE: under the current version of MX (3.3 and below), the queue file X$! will be open if any user is in the process of sending an MX mail message. V X$! This is because MX opens the file before the user is asked to supply the X$! message and keeps it open throughout the entry of that \3707. X$! X$! --------\0808\1010\2020\400C X$ Tries = 0 X$ wait_loop: X$ if Tries .gt. Max\0E06then $ goto Wait_Too_Long X$ Tries =\0807+ 1 X$ wait 'Delay' X$ open/read/error=wait_loop tmp 'MXqueue' X$ close tmp X$ X$! --------\0808\1010\2020\400C X$! Now we can do the compress;\0A09 down to a temporary file first and X$! then rename it to the master one. This prevents unexpected aborts from X$! leaving partial queue files. X$! X$! NOTE: X$!`09While we are compressing, we also optimize the FDL file based on the X$!`09usage of the queue file. X$! --------\0808\1010\2020\400C X$ em "-I-ANALYZE, Analyzing system queue FDL" X$ analyze/rms/fdl/output='MXfdl'_temp 'MXqueue' X$ if .not. $status X$ then X$ em "-I-STATUS, ", $status, " : ", f$message(\1A07) X$ em "-E-ANALFAIL, Analysis of system queue failed - aborted" X$ delete 'MXfdl'_temp;* X$ goto CompressNotDone X$ endif X$ X$ em "-I-OPTIMIZE, Optimizing the system queue FDL" X$ edit/fdl/nointeractive/analysis='MXfdl'_temp \0D0C X$ if .not. $status X$ then X$ em "-I-STATUS, ", $status, " : ", f$message(\1A07) X$ em "-E-OPTIFAIL, Optimisation of system queue failed - aborted" X$ delete 'MXfdl'_temp;* X$ goto CompressNotDone X$ endif X$ X$ X$ em "-I-COMPRESS, Compressing MX queue file using optimised FDL" X$ convert/nosort - X /fdl='MXfdl'_temp - X`09'MXqueue' - X 'MXqueue'_temp X$ if .not. $status X$ then \0606\0C0C\180A7 X$ em "-I-STATUS, ", $status, " : ", f$message(\1A07) X$ em "-I-COMPFAIL, Unable to complete the queue compress - aborted" X$ delete 'MXqueue'_temp;* X$ delete 'MXfdl'_temp;* X$ goto CompressNotDone X$ endif X$ X$! --------\0808\1010\2020\400C X$! Rename the newly built queue file to the master\1306and purge off old X$! copies X$! --------\0808\1010\2020\400C X$ em "-I-UPDATE, Updating queue file" X$ rename 'MXqueue'_temp\0F0A X$ purge 'MXqueue' X$ rename 'MXqueue' ;1 X$ X$ delete 'MXfdl'_temp;* X$ X$ X$! --------\0808\1010\2020\400C X$! Compress aborted, finished or timed out; restart MX X$! X$! The restart is complicated by two things. First, if the system is in a X$! cluster and the\1009mode is selected, it is necessary to restart MX X$! on all the nodes in the cluster that originally ran it. We do this by using V X$! either SYSMAN or a set of node specific BATCH queues. Second, this X$! procedure may be run by any user with the right privileges; however, MX mus Vt X$! be restarted under it's original username (see p4). If the requires\2309 X$! is the same as our current one, we use SYSMAN otherwise we submit the X$! startup procedure to each of the specified batch queues (see P1). X$! X$! --------\0808\1010\2020\400C X$CompressNotDone: X$Wait_Too_Long: X$RestartMx: X$ if f$edit(f$getjpi("","USERNAME"),"UPCASE,TRIM") .eqs. Username X$ then X$ X$ ! Issue the restart command on each node in the cluster using the sysman X$ ! command. Note that the MX_ROOT:MX_STARTUP_INFO.DAT file should X$ ! specifically name those nodes on which MX components should run because X$ ! this procedure has no knowledge of which nodes in the cluster run MX X$ ! and which don't. X$ em "-I-RESTART, Restarting MX on all nodes via SYSMAN" X$ mc sysman set env/cluster X set timeout 00:00:30`09`09&0209! In case of hangs! X do @sys$startup:mx_\0B07`09`09`09! Restart cluster-wide X$ else X$ X$ ! Start up the restart job by submitting it to a batch queue appropriate to V X$ ! the node (see p1) X$ n=0 X$ QueueSubmit: X$ q = f$element(n, ",", Queues) X$ if q .nes. "," X$ then X$ x = f$getqui("CANCEL_OPERATION") X$ node = f$getqui("DISPLAY_QUEUE", "SCSNODE_NAME", q, "ALL_JOBS") X$ em "-I-RESTART, Restarting MX on node ", node, " via job on batch q Vueue ", q X$ submit/user='Username'/nolog/queue='q' sys$startup:mx_\0B07 X$ n=n+1 X$ goto QueueSubmit X$ endif X$ endif $ call unpack MX_SHRINK.COM;17 90116477 "" 20 1 2 $! $ create 'f' XMX QUEUE FILE COMPRESSION X X8-NOV-1993 X XAndy Harper`09`09&0209&0309Version 1.4 X X--------\0808\1010\2020\400F X X XINTRODUCTION X XThe queue file used by the MX mail system (MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL) Xwill, on a heavily used system, grow unbounded. This causes performance Xproblems due to excessive use of file extent pointers (hence lots of disk Xseeks). X XThe DCL procedure described here is an attempt to reduce the performance Xproblems by periodically compressing the queue file. X X X XOPERATION X XThe queue file cannot be compressed while MX is running (strictly,\1F07any Xuser has the file open) so MX needs to be shut down. The procedure will first V Xshut MX down, then compress the queue file, and finally restart MX. Depending V Xon the size of the queue file, MX may be unavailable for some considerable Xtime. It is recommended that this procedure be run at the quietest moment Xovernight. X XThe procedure attempts to be fail safe. When the shutdown signal has been sen Vt Xto MX, it loops trying to open the queue file. Once it can, no-one is accessin Vg Xit and this indicates that MX has completely shut down (after finishing Xwhatever it was doing in a tidy manner). However, should the procedure be Xunable to open the queue file after a number of retries, it will give up and Xrestart MX. X XMX is also restarted if the procedure is aborted or\1F08compression attempt Xfails in some way. X XOnce the queue is accessible to the procedure, the compression operation Xbegins. A new file is created using the currently valid records of the existin Vg Xqueue. This new file is given a different name (and later renamed) to avoid a V Xpartial queue file being inadvertently put in place as the master. X XThe OPTIMIZE script of EDIT/FDL is used to produce an optimal FDL file for the V Xsystem queue. This helps to ensure optimum performance. X X X X XRUNNING THE PROCEDURE X XThe procedure is run as follows: X X @MX_SHRINK `5Bbatch-queues`5D `5Bdelay`5D `5Bretry`5D `5Busername`5D X Xwhere: X X `5Bbatch-queues`5D X`09`09Specifies a comma separated list of batch queues into which X`09`09the MX restart job is submitted. Each queue should execute on X`09`09one of the nodes in the cluster and there should be one queue X`09`09for each node on which MX runs. X X`09`09If the node is stand-alone, then this paramter is OPTIONAL X`09`09and defaults to SYS$BATCH. In a cluster, a suitable list X`09`09of batch queues is required. These must be \2806execution X`09`09queues. X X`09`09Note this parameter is NOT used if the procedure is submitted X`09`09under the username given by the `5B\1708`5D parameter. Instead, X`09`09the procedure is submitted using SYSMAN. X X `5Bdelay`5D`09is the delta time to wait between retries `5Bdefault: 15 seco Vnds`5D X X `5Bretry`5D`09is the number of times to retry`09if the queue file cannot be V X`09`09opened (default: 240) X X `5Busername`5D`09is the name under which to restart MX `5Bdefault: SYSTEM V`5D X X XYou should arrange to run this procedure on a regular basis every night by som Ve Xexternal means. X X X XACKNOWLEDGEMENTS X XI'm grateful to Hunter Goatley, who provided the basic draft of this procedure V, Xfor permission to make his code freely available in my modified form. X`0C XCHANGES FROM VERSION 1.3 X X1.`09Added automatic optimization of the queue file characteristics, using X`09the OPTIMIZE script of EDIT/FDL. X X X XCHANGES FROM VERSION 1.2 X X1.`09SETPRV privilege is needed to startup MX; we now check for this. X X2.`09Various typos fixed. X X3.`09Added /NOSORT to the CONVERT/FDL command to speed things up a bit X X4.`09Added a QUEUE PURGE command prior to shutting everything down. X X X XCHANGES FROM VERSION 1.1 X X1.`09SYSLCK privilege is required for an MX shutdown and SETPRV\2A0C X`09to restart MX; this is now checked before any actions are taken. X X2.`09The procedure now tries to deal with MX running in a cluster and using X`09a shared MX queue file by: X `09`09(a)`09Issuing a SHUTDOWN /CLUSTER command via MCP X`09`09(b)`09Restarting MX by submitting the MX_STARTUP job either X`09`09`09via SYSMAN or into a list of node specific batch queues X`09`09`09supplied as parameters. X X`09As a consequence, the parameter list of\1606rocedure has changed. X X`09See the documentation and source code for full details of how this X`09works. X X3.`09Fixed a typo in the documentation that transposed the parameters X`09`5Bretry`5D and `5Bdelay`5D. $ call unpack MX_SHRINK.DOC;6 1197764331 "" 9 2 2 $ v=f$verify(v) $ exit ================================================================================ Archive-Date: Tue, 09 Nov 1993 11:51:35 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: Tuning MX V3.3 queue file Message-ID: <1993Nov9.102048.309@buckie.hsc.colorado.edu> From: Date: 9 Nov 93 10:20:47 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Comments welcome. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ----- Tuning your MX queue file MX V3.3 Dan Wing dwing@uh01.colorado.edu Before you use Andy Harper's new MX_SHRINK, you need to tune MX's queue file on your system. The exact values necessary for the file, of course, are site-specific (ever notice that everything important is "site specific"?), but the following procedure should get you reasonably close to a 'good' file: NOTE: Performing these steps requires sufficient privileges to shutdown and restart MX -- SETPRV is recommended to ensure you have sufficient privileges to restart MX. 1. Shutdown all MX agents -- this is to get all processes to let go of the queue file: $ MCP := $MX_EXE:MCP.EXE $ MCP SHUTDOWN/CLUSTER 2. Generate an FDL (File Description Language) file which describes the current file. The FDL file is a plain text file. The ANALYZE/RMS/FDL command will create the file SYSTEM_QUEUE.FDL. This command could take 10 seconds to 10 minutes, depending on the file's size and degree of fragmentation. $ SET DEFAULT MX_FLQ_DIR: $ ANALYZE/RMS/FDL MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL 3. If you want, you can startup all MX agents at this time, but you'll have to shut them down to actually tune the queue file in a few steps.... 4. Now we're going to tune the FDL file. This doesn't tune the queue file itself, it just prepares us for the tuning (or for tunning Andy's MX_SHRINK.COM program). There are two methods, the "Quick and Dirty Method" and the "Long, Boring Method": Quick and Dirty Method ---------------------- 4a. Have the FDL editor do all the work: $ SET DEFAULT MX_FLQ_DIR: $ EDIT/FDL/ANALYZE=SYSTEM_QUEUE.FDL/NoINTERACTIVE SYSTEM_QUEUE.FDL Long, Boring Method ------------------- 4a. Fire up the FDL editor (we turn off ReGis so you don't get the somewhat slow and useless ReGis displays from EDIT/FDL if you're running in a DECterm): $ SET TERMINAL/NoREGIS $ SET DEFAULT MX_FLQ_DIR: $ EDIT/FDL/ANALYZE=SYSTEM_QUEUE.FDL SYSTEM_QUEUE.FDL 4b. You'll be presented with a screen. Type: I (Invoke) O (Optimize) L (Line graph) 4c. "Number of Records that will be Reloaded": (to accept the default value) 4d. "File Reloading Method": will be doing a Fast reload. (to accept the default value [Fast]) 4e. "Number of Additional Records to be Added After the Reloading [of] the File": A reasonable number here, such as the number of MX messages you process each day * 30, would be a good start: 30000 4f. "Will Additional Records Typically be Added in Order by Ascending Primary Key". The primary key is the MX queue number, which is incremented by one for each message. YES 4g. "Will Added Records be Distributed Evenly over the Reloaded Range of Pri Key Values": No, they will not be evenly distributed, they will all be at the end (*): NO 4h. You will see a graph, and then the prompt "While File Parameter": FD (Finish Design) 4i. "Text for FDL Title Section": You can enter whatever you like; your name (or username) and the date you performed the tune might be useful. 4j. You'll see a comparison of using different bucket sizes, and the prompt "Key 0 Bucket Size": Press RETURN to accept the default value. <RETURN> 4k. The FDL editor will then spit out an informational message about the depth of key 0, and other information. Press RETURN to continue. <RETURN> 4l. Back at the main menu. At the prompt "Main Editor Function" type EXIT: EXIT (to leave the FDL Editor after creating the FDL file) (*) Note: Eventually, MX's queue numbering will restart at message number 1, so you could argue that the keys are evenly distributed. How soon this happens depends on how many messages are processed by your site. 5. You now have a "tuned" .FDL file which can be used by Andy's MX_SHRINK.COM program, or you can perform the convert by hand: 5a. Shutdown all MX agents: $ MCP := $MX_EXE:MCP.EXE $ MCP SHUTDOWN/CLUSTER 5b. Perform conversion: $ SET DEFAULT MX_FLQ_DIR: $ RENAME SYSTEM_QUEUE.FLQ_CTL SYSTEM_QUEUE.LOCKED $ CONVERT/noSORT/FDL=SYSTEM_QUEUE.FDL/STATISTICS - SYSTEM_QUEUE.LOCKED - SYSTEM_QUEUE.TMP $ RENAME/LOG SYSTEM_QUEUE.TMP SYSTEM_QUEUE.FLQ_CTL $ DELETE/LOG SYSTEM_QUEUE.LOCKED; If you receive a "file locked" message when you perform the CONVERT step, a user on your system opened the queue file before you RENAMEd it (most likely the user is composing a message to be sent). ================================================================================ Archive-Date: Tue, 09 Nov 1993 13:55:45 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 09 Nov 1993 14:05:34 EDT From: Dan Wheeler <manager@LEMOYNE.BITNET> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-LIST@WKUVX1.WKU.EDU CC: manager@MAPLE Message-ID: <00975491.F5339AA0.6627@lemoyne> Subject: Local .TMP files question I just discovered what I presume is a problem with my MX 3.3/VMS 5.5-2 configuration. While attempting to start a debug of LOCAL for another problem I discovered that I had *tons* (I never had the patience to wait over 10 minutes for completion of a DIR/TOTAL) of .TMP files in the [.LOCAL] directory, some of them more than 2 months old. The directory file itself is >3000 blocks and it is taking me hours just to delete the old files. My local accounting is enabled. I haven't found anything in the docs or troubleshooting guide that helps me yet. Where should I start looking? --dan --------------------------------------------------------------------------- Dan Wheeler Senior Support Analyst & VMS System Manager Office of Information Systems, Le Moyne College, Syracuse NY 13214-1399 Phone 315.445.4565 (Home 315.655.8193) WHEELER@LEMOYNE.BITNET --------------------------------------------------------------------------- ================================================================================ Archive-Date: Tue, 09 Nov 1993 14:36:59 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Local .TMP files question Message-ID: <1993Nov9.132303.312@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 9 Nov 93 13:23:03 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17334176@MVB.SAIC.COM>, Dan Wheeler <manager@LEMOYNE.BITNET> writes: >I just discovered what I presume is a problem with my MX 3.3/VMS 5.5-2 >configuration. While attempting to start a debug of LOCAL for another >problem I discovered that I had *tons* (I never had the patience to wait >over 10 minutes for completion of a DIR/TOTAL) of .TMP files in the >[.LOCAL] directory, some of them more than 2 months old. The directory >file itself is >3000 blocks and it is taking me hours just to delete the >old files. My local accounting is enabled. > >I haven't found anything in the docs or troubleshooting guide that helps >me yet. Where should I start looking? I'm sure the file MX_ROOT:[00000]LOCAL.DIR has exceeded 128 blocks, so you should really rebuild the .DIR file so VMS can cache it (to rebuild it, you can just shutdown all MX agents, move all files out of the directory, delete the directory, re-create the directory, and then move the files back into the directory.) The only thing I could guess that caused the .TMP files is a failed local delivery which, for some reason, didn't cause a bounce (I've had a few failed local deliveries when my MAIL.MAI file is locked [when I'm receiving lots of Email all at the same time], but I haven't been able to reliably duplicate it). -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Tue, 09 Nov 1993 14:57:35 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 09 Nov 1993 14:26:59 CST From: "John Clement-Rice U. (713)527-8101x2037" <clement@physics.rice.edu> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975494.F38239C0.18073@physics.rice.edu> Subject: Re: Missing Date: field ? My thanks to the various people who sent me a reply about this problem. The earliest reply and the winner was David Hastings. I am putting this onto the list as I am sure that others have or will run into the same problem. >>We are having a problem with mail that is sent by Eudora on a Mac, >>through MX. Essentialy Eudora is leaving off the Date: field. >>When I get a message from Eudora it does not have the field, while any message >>I send using MX or virtually any other mailer seems to have a Date field. >> >>1. Does anyone know how to coerce Eudora into putting on the date??? > >Yes. You need to run the Map control panel to on the Mac set your timezone. This >is very obscure but it does work! > >Dave >-- > >David Hastings | "There's nothing wrong with my sense >VAX Systems Programmer | of reality - I have it thoroughly >Oxford University Computing Services| serviced once a fortnight" - DNA > Now for those who do not know how to do this bring up the map control panel in Mac System 7. Point to your location and click. Then point to the SET button and click. --------------> Editorial comment <---------------- Now I can not for the life of me understand why some mailers bother to require the Date: field, on incoming mail. MX being reasonable does not care. Also why in the world does Eudora really care about the timezone? Here is an example of an application, that does not obey the standards. It should put the date into the message whether or not it knows the timezone. It should be noted that the older version of Eudora worked just fine without the timezone. It should at least warn you that you need to set the timezone, or refuse to send mail until it knows the timezone. Having vented my spleen, I would like to again thank all of the kind folks without whom, I would be up the creek. QUIT |/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ | John M. Clement Hepnet: UTSPAN::RIPHYS::CLEMENT | Bonner Lab Bitnet: CLEMENT@RICEVM1 | Rice University Internet: CLEMENT@PHYSICS.RICE.EDU | PO Box 1892 BellTel: (713)527-8181 x2037 | Houston, Tx. 77251 Fax: (713)285-5215 |/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ ================================================================================ Archive-Date: Tue, 09 Nov 1993 17:18:25 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: No outgoing mail. help. Message-ID: <2bp77gINNaec@jhunix.hcf.jhu.edu> From: nils@jhunix.hcf.jhu.edu (Nils Gokemeijer) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 9 Nov 1993 17:56:48 -0500 To: MX-List@WKUVX1.WKU.EDU I just installed mx, while using cmu-tek v6.4. EVerything seems fine, and it all works, except: Nobody can send any mail over internet except for system. system can do anything no problem. but if a regular user mails something, it just vanishes into nowhere. I assume it is some priviledge problem, but I can't figure it out. Here are two log files from mx_router. btw: I am running under system, no separate mail-account 9-NOV-1993 12:22:09.64 %PROCESS, Processing entry number 214 9-NOV-1993 12:22:09.97 %PROCESS, Status from READ_INFO was 00000001 9-NOV-1993 12:22:09.97 %PROCESS, Message originated in VMS Mail. 9-NOV-1993 12:22:10.01 %PROCESS, will run domain expander on envelope addresses. 9-NOV-1993 12:22:10.02 %PROCESS, will run domain expander on message headers. 9-NOV-1993 12:22:12.41 %PROCESS, Finished VMSmail-origin preprocessing. 9-NOV-1993 12:22:12.42 %PROCESS, Marking this entry as finished. You can see it does not do anything! (does not expand envelope address!) Now this second one runs fine: (expanding works) can someone tell me the problem? 9-NOV-1993 12:53:40.39 %PROCESS, Processing entry number 215 9-NOV-1993 12:53:40.51 %PROCESS, Status from READ_INFO was 00000001 9-NOV-1993 12:53:40.52 %PROCESS, Message originated in VMS Mail. 9-NOV-1993 12:53:40.56 %PROCESS, will run domain expander on envelope addresses. 9-NOV-1993 12:53:40.57 %PROCESS, Processing address: <nils@eta.pha.jhu.edu> 9-NOV-1993 12:53:40.64 %PROCESS, ... address now reads: <nils@eta.pha.jhu.edu> 9-NOV-1993 12:53:40.65 %PROCESS, will run domain expander on message headers. 9-NOV-1993 12:53:41.68 %PROCESS, Finished VMSmail-origin preprocessing. 9-NOV-1993 12:53:41.69 %PROCESS, Recipient #0: <nils@eta.pha.jhu.edu> 9-NOV-1993 12:53:41.72 %REWRITE, No rewrite rules matched <nils@eta.pha.jhu.edu> 9-NOV-1993 12:53:41.89 %FINDPATH, Site-spec expand on eta.pha.jhu.edu err=00000000 9-NOV-1993 12:53:41.90 %FINDPATH, domain name ETA.PHA.JHU.EDU matched path pattern * 9-NOV-1993 12:53:41.91 %PROCESS, Rewrote <nils@eta.pha.jhu.edu> as <nils@eta.pha.jhu.edu> - next hop eta.pha.jhu.edu, path 2 9-NOV-1993 12:53:42.24 %PROCESS, Adding to SMTP path: <nils@eta.pha.jhu.edu>. 9-NOV-1993 12:53:43.19 %PROCESS, Path SMTP gets 1 rcpts, entry number 216 Thanks, Nils ================================================================================ Archive-Date: Tue, 09 Nov 1993 17:53:06 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: No outgoing mail. help. Message-ID: <1993Nov9.162147.313@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 9 Nov 93 16:21:47 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <2bp77gINNaec@jhunix.hcf.jhu.edu>, nils@jhunix.hcf.jhu.edu (Nils Gokemeijer) writes: >I just installed mx, while using cmu-tek v6.4. EVerything seems fine, >and it all works, except: Nobody can send any mail over internet >except for system. system can do anything no problem. but if a regular >user mails something, it just vanishes into nowhere. I assume it is >some priviledge problem, but I can't figure it out. Here are two log >files from mx_router. >btw: I am running under system, no separate mail-account What version of VMS are you running, what version of MX are you running? Do you have security auditing enabled so you get reports of file access failures? Are there any file access failures when a user tries to send mail? Does incoming mail work at all? Are you going to get my CC'd mail message I'm sending to you? -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Wed, 10 Nov 1993 06:12:31 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 10 Nov 1993 06:11:52 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: udaa055@BAY.CC.KCL.AC.UK Message-ID: <00975518.F2F73F67.18768@ALPHA.WKU.EDU> Subject: RE: MX queue compression - MX_SHRINK V1.4 "Andy, Systems Manager" <udaa055@BAY.CC.KCL.AC.UK> writes: > >I've improved the MX_SHRINK utility, recently posted, by adding some code >to optimize the size of the MX queue file. This is handled by using the >optimize script of the EDIT/FDL command. > Thanks, Andy! I used it last night in place of my stupid version. For anybody planning to use MX_SHRINK V1.4, I found a typo: $ Username = P3 $ if Username .eqs. "" then $ Username = "SYSTEM" ! Username for MX startup The first line should read: $ Username = P4 Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Wed, 10 Nov 1993 08:11:11 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 10 Nov 1993 09:00:37 EST From: "Brian Tillman" <tillman@swdev.si.com> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wku.edu Message-ID: <00975530.86193860.13859@swdev.si.com> Subject: MX_SHRINK V1.4 comment I unpacked Andy's MX_SHRINK V1.4 and I came across the following line: $ convert/nosort - /fdl='MXfdl'_temp - 'MXqueue' - 'MXqueue'_temp $ if .not. $status $ then 7 $ em "-I-STATUS, ", $status, " : ", f$message($status) $ em "-I-COMPFAIL, Unable to complete the queue compress - aborted" $ delete 'MXqueue'_temp;* $ delete 'MXfdl'_temp;* $ goto CompressNotDone $ endif Notice the "7" hanging out there on the THEN statement in the middle. I checked the SHARE file. This is what's on that line: X$ then \0606\0C0C\180A7 Somehow, this doesn't seem correct. -- Brian Tillman tillman_brian@si.com ================================================================================ Archive-Date: Wed, 10 Nov 1993 08:41:18 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Tuning MX V3.3 queue file Message-ID: <1993Nov10.121349.18075@vax.oxford.ac.uk> From: daveh@vax.oxford.ac.uk (Dave Hastings, OUCS) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 10 Nov 93 12:13:49 GMT To: MX-List@WKUVX1.WKU.EDU > Tuning your MX queue file > MX V3.3 > > Dan Wing > dwing@uh01.colorado.edu > > > Long, Boring Method > ------------------- > > 4j. You'll see a comparison of using different bucket sizes, and the > prompt "Key 0 Bucket Size": Press RETURN to accept the default > value. > <RETURN> MX seems to work better if the queue file's cluster size is divisible by the bucket size (On our system the cluster size is 51 blocks and the bucket size is 17 blocks). > > 4k. The FDL editor will then spit out an informational message > about the depth of key 0, and other information. Press RETURN > to continue. > <RETURN> The index should be as flat as possible. Dave -- David Hastings | "There's nothing wrong with my sense VAX Systems Programmer | of reality - I have it thoroughly Oxford University Computing Services| serviced once a fortnight" - DNA daveh@vax.oxford.ac.uk | ================================================================================ Archive-Date: Wed, 10 Nov 1993 09:51:20 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 10 Nov 1993 07:10:50 PDT From: Jim Wheeler # 916-978-5089 <jimww@sacto.mp.usbr.gov> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu CC: jimww@sacto.mp.usbr.gov Message-ID: <00975521.2FFEA406.30016@sacto.mp.usbr.gov> Subject: Listserver question howdy, I have accounts on several machines and want to be able to post from them but receive mail on only one. To that end I added myself from two other nodes, one through a wpo/smtp gateway. The wpo generated an uppercase userid so I used the nocase switch, but when I tried to post mx required me to resubscirbe. This is after I did. Please, note the nocase in the third entry and that the fourth is identical, except for case, as the third. What am I doing wrong? jimww@SACTO.MP.USBR.GOV jwheeler@MPNETMGR.MP.USBR.GOV (NOMAIL,NOCASE,CONCEAL,NOREPRO) jwwheeler%ibr2smtp@MPNETMGR.MP.USBR.GOV (NOMAIL,NOCASE,CONCEAL,NOREPRO) JWWHEELER%ibr2smtp@MPNETMGR.MP.USBR.GOV HELP!!!! quit ================================================================================ # James W. Wheeler # # # Regional Systems & WAN Manager # voice: 916 978-5089 fax: 916-978-5284 # # Mid Pacific Region # # # U.S. Bureau of Reclamation # inet: jimww@sacto.mp.usbr.gov # # 2800 Cottage Way MP1130 # # # Sacramento, CA 95825 # # ================================================================================ ================================================================================ Archive-Date: Wed, 10 Nov 1993 10:13:12 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 10 Nov 1993 15:10:44 EST From: "Andy, Systems Manager" <udaa055@bay.cc.kcl.ac.uk> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: udaa055@bay.cc.kcl.ac.uk Message-ID: <00975564.3A24F460.16503@bay.cc.kcl.ac.uk> Subject: RE: MX_SHRINK V1.4 comment Brian Tillman writes: >Notice the "7" hanging out there on the THEN statement in the middle. I checked >the SHARE file. This is what's on that line: >X$ then \0606\0C0C\180A7 > >Somehow, this doesn't seem correct. Brian, You are quite right; this 7 is in fact in the original code. It doesn't seem to affect operation unless there's an error causing the if branch to be taken; but, of course, it should not be there. It was a typo caused by a late night and insufficient coffee! I have fixed this in 1.5, together with a few other things and I hope to distribute this in the next day or so. Regards, Andy Harper Kings College London ================================================================================ Archive-Date: Wed, 10 Nov 1993 11:43:22 CST Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 10 Nov 1993 17:40:04 EST From: "Andy, Systems Manager" <udaa055@bay.cc.kcl.ac.uk> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <00975579.1699EFE0.16981@bay.cc.kcl.ac.uk> Subject: MX_SHRINK 1.5 Dear All, Thank you for your helpful comments on the MX_SHRINK procedure recently posted. Enclosed is the latest version which fixes a few typos and sorts out a minor timing window. If no bugs are found, I plan to make this the last version. Regards, Andy Harper Kings College London $! ------------------ CUT HERE ----------------------- $ v='f$verify(f$trnlnm("SHARE_UNPACK_VERIFY"))' $! $! This archive created: $! Name : MX_SHRINK $! By : Andy Harper <udaa055@bay.cc.kcl.ac.uk> $! Date : 10-NOV-1993 17:34:03.65 $! Using: VMS_SHARE 8.4, (C) 1993 Andy Harper, Kings College London UK $! $! Credit is due to these people for their original ideas: $! James Gray, Michael Bednarek $! $! TO UNPACK THIS SHARE FILE, CONCATENATE ALL PARTS IN ORDER $! AND EXECUTE AS A COMMAND PROCEDURE ( @name ) $! $! THE FOLLOWING FILE(S) WILL BE CREATED AFTER UNPACKING: $! 1. MX_SHRINK.COM;24 $! 2. MX_SHRINK.DOC;10 $! $ set="set" $ set symbol/scope=(nolocal,noglobal) $ f=f$parse("SHARE_UNPACK_TEMP","SYS$SCRATCH:."+f$getjpi("","PID")) $ e="write sys$error ""%UNPACK"", " $ w="write sys$output ""%UNPACK"", " $ if .not. f$trnlnm("SHARE_UNPACK_LOG") then $ w = "!" $ if f$getsyi("CPU") .gt. 127 then $ goto start $ ve=f$getsyi("version") $ if ve-f$extract(0,1,ve) .ges. "4.4" then $ goto start $ e "-E-OLDVER, Must run at least VMS 4.4" $ v=f$verify(v) $ exit 44 $unpack: subroutine ! P1=filename, P2=checksum, P3=attributes,P4=size $ if f$parse(P1) .nes. "" then $ goto dirok $ dn=f$parse(P1,,,"DIRECTORY") $ w "-I-CREDIR, Creating directory ''dn'" $ create/dir 'dn' $ if $status then $ goto dirok $ e "-E-CREDIRFAIL, Unable to create ''dn' File skipped" $ delete 'f'* $ exit $dirok: $ x=f$search(P1) $ if x .eqs. "" then $ goto file_absent $ e "-W-EXISTS, File ''P1' exists. Skipped" $ delete 'f'* $ exit $file_absent: $ w "-I-UNPACK, Unpacking ", P5, " of ", P6, " - ", P1, " - ", P4, " Blocks" $ n=P1 $ if P3 .nes. "" then $ n=f $ if .not. f$verify() then $ define/user sys$output nl: $ EDIT/TPU/NOSEC/NODIS/COM=SYS$INPUT/NOJOURNAL 'f'/OUT='n' PROCEDURE GetHex(s,p)LOCAL x1,x2;x1:=INDEX(t,SUBSTR(s,p,1))-1;x2:=INDEX(t, SUBSTR(s,p+1,1))-1;RETURN 16*x1+x2;ENDPROCEDURE;PROCEDURE SkipPartsep LOCAL m; LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m);EXITIF INDEX(ERASE_LINE, "-+-+-+-+-+-+-+-+")=1;ENDLOOP;ENDPROCEDURE;PROCEDURE ProcessLine LOCAL c,s,l,b, n,p;c := ERASE_CHARACTER(1);s := ERASE_LINE;EDIT(s,"TRIM");IF c = "X" THEN SPLIT_LINE; ENDIF;MOVE_HORIZONTAL(-1);l := LENGTH(s);p := 1; LOOP EXITIF p > l;c := SUBSTR(s,p,1);p := p+1;CASE c FROM ' ' TO '`' ['\']: b:= GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT( SUBSTR(CURRENT_LINE, CURRENT_OFFSET-b+1,n));['&']: b:=GetHex(s,p); n:=GetHex(s,p+2); p:=p+4; COPY_TEXT(ASCII(n)*b);['`']: COPY_TEXT(ASCII(GetHex(s,p))); p:=p+2;[INRANGE, OUTRANGE]: COPY_TEXT(c);ENDCASE;ENDLOOP;ENDPROCEDURE;PROCEDURE Decode LOCAL m; POSITION(BEGINNING_OF(b));LOOP m:=MARK(NONE);EXITIF m=END_OF(b);DELETE(m); IF INDEX(CURRENT_LINE,"+-+-+-+-+-+-+-+-")=1 THEN SkipPartSep;ELSE ProcessLine; MOVE_HORIZONTAL(1);ENDIF;ENDLOOP;ENDPROCEDURE;SET(FACILITY_NAME,"UNPACK");SET( SUCCESS,OFF);SET(INFORMATIONAL,OFF);t:="0123456789ABCDEF";f:=GET_INFO( COMMAND_LINE,"file_name");b:=CREATE_BUFFER(f,f);Decode;WRITE_FILE(b,GET_INFO( COMMAND_LINE,"output_file"));QUIT; $ if p3 .eqs. "" then $ goto dl $ open/write fdl &f $ write fdl "RECORD" $ write fdl P3 $ close fdl $ w "-I-CONVRFM, Converting record format to ", P3 $ convert/fdl=&f &f-1 &P1 $dl: delete 'f'* $ checksum 'P1' $ if checksum$checksum .nes. P2 then $ - e "-E-CHKSMFAIL, Checksum of ''P1' failed." $ exit $ endsubroutine $start: $! $ create 'f' X$! Purpose: X$!`09Automatically compress the MX QUEUE file. X$! X$! Parameters: X$!`09P1`09List of queues for restart job. if a stand-alone node, this X$!`09`09is optional and defaults to SYS$BATCH. If a member of a cluster X$!`09`09a list of execution batch queues (one per node) is required X$!`09`09This parameter is ignored if P4 (see below) is the same as the X$!`09`09current user. SYSMAN is used to restart MX in this case. X$!`09P2`09`5Boptional`5D delay time between queue file access retries (default V: 15 seconds) X$!`09P3`09`5Boptional`5D number of retries (default: 240 ) X$!`09P4`09`5Boptional`5D username to run MX (default: SYSTEM) X$! X$! Privileges: X$!`09SYSPRV X$!`09CMKRNL X$!`09SYSLCK X$!`09SETPRV X$!`09 X$! Environment: X$!`09Requires MX logicals to be defined X$!`09Requires MX software to be present X$! X$! Revision History: X$!`091.0`09??????????`09Hunter Goatley`09Original version X$!`091.1`091-NOV-1993`09Andy Harper`09Added parameters, comments, error checki Vng X$!`091.2`095-NOV-1993`09Andy Harper`09Add cluster support; fix privs X$!`091.3`098-NOV-1993`09Andy Harper`09Minor typos; add queue purge X$!`091.4`098-NOV-1993`09Andy Harper`09Add automatic optimization X$!`091.5`0910-NOV-993`09Andy Harper`09Move QUEUE PURGE after FDL opt X$!`09`09&0209&0209`26 rejig queue access timing X$! X$ X$ On control_y then $ goto RestartMX X$ set noon X$ em = "write sys$output ""%"+f$parse(f$environment("PROCEDURE"),,,"NAME")+""" V," X$ mcp = "$mx_exe:mcp.exe/file=mx_dir:mx_config.mxcfg" X$ Cluster = f$getsyi("CLUSTER_MEMBER") X$ X$ MXqueue = "MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL" X$ MXfdl = "MX_FLQ_DIR:SYSTEM_QUEUE.FDL" X$ RMS$_FLK= %X10B2109C`09`09&0209! File locked by another user X$ X$ X$! --------\0808\1010\2020\400C X$! Set up the parameters; default if necessary X$! --------\0808\1010\2020\400C X$ Queues = P1 X$ if Queues .eqs. "" then $ \1909= "SYS$BATCH"`09! Batch queue for startup X$ Delay = P2 X$ if Delay .eqs. "" then $ \1909= "00:00:15"`09! Interval between tries X$ MaxTries = P3 X$ if MaxTries .eqs. "" then $ \1909= 240`09`09! Number of retries; an hour max Vimum X$ Username = P4 X$ if Username .eqs. "" then $ \1909= "SYSTEM"`09!\2F0Afor MX startup X$ X$ X$! --------\0808\1010\2020\400C X$! Check privileges required X$! --------\0808\1010\2020\400C X$ if .not. f$priviledge("SYSPRV,CMKRNL,SYSLCK") X$ then X$ em "-E-INSUFPRV, ", f$fao("!8%T : Insufficient privilege",0) X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Make sure we have access to the MCP program so we can shut down everything V X$! --------\0808\1010\2020\400C X$ if f$search("mx_exe:mcp.exe") .eqs. "" X$ then X$ em "-E-NOMCP, ", f$fao("!8%T : MX control program MCP not found",0) X$ exit X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Check that we have a valid list of batch queue if CLUSTER operation has bee Vn X$! specified. The batch queues must execute on the nodes on which MX runs. X$! --------\0808\1010\2020\400C X$ if Cluster .eqs. "" X$ then X$ if P1 .eqs. "" X$ then X$ em "-E-CLUBATQUE, ", f$fao("!8%T : Batch execution queue list needed; V member of cluster",0) X$ exit X$ endif X$ endif X$ X$! --------\0808\1010\2020\400C X$! Check the list of queues supplied to ensure that they are BATCH\2E08and X$! that they are EXECUTION queues rather than GENERIC\1B08- we want the X$! startup to execute on known nodes. X$! --------\0808\1010\2020\400C X$ n=0 X$QueueCheck: X$ q = f$element(n, ",", Queues) X$ if q .nes. "," X$ then X$ x = f$getqui("CANCEL_OPERATION") X$ x = f$getqui("DISPLAY_QUEUE", "QUEUE_NAME", q) X$ if x .eqs. "" X$ then X$ em "-E-QUENONEXI, ", f$fao("!8%T : !AS : not a valid queue name",0,q V) X$ exit X$ endif X$ if .not. f$getqui("DISPLAY_QUEUE", "QUEUE_BATCH", q) X$ then X$ em "-E-QUENOTBAT, ", f$fao("!8%T : !AS : not a batch queue",0,q) X$ exit X$ endif X$ if f$getqui("DISPLAY_QUEUE", "QUEUE_GENERIC", q) X$ then X$ em "-E-QUENOTEXE, ", f$fao("!8%T : !AS : must be execution queue not V generic",0,q) X$ exit X$ endif X$ n=n+1 X$ goto QueueCheck X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Shut all the MX processes having the queue file open X$! --------\0808\1010\2020\400C X$ if Cluster X$ then X$ em "-I-SHUTDOWN, ", f$fao("!8%T : Shutting down MX cluster wide",0) X$ mcp shutdown /cluster X$ else X$ em "-I-SHUTDOWN, ", f$fao("!8%T : Shutting down MX",0) X$ mcp shutdown X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Loop until we can OPEN the SYSTEM_QUEUE file; when \2806, nobody X$! else is using it, which means all the agents have shutdown. X$! However, to avoid MX being shutdown forever due to any system problems X$! elsewhere, limit the amount of time we retry (see P2 and p3). X$! X$! NOTE: under the current version of MX (3.3 and below), the queue file X$! will be open if any user is in the process of sending an MX mail message. V X$! This is because MX opens the file before the user is asked to supply the X$! message and keeps it open throughout the entry of that \3707. X$! X$! --------\0808\1010\2020\400C X$ Tries = 0 X$ wait_loop: X$ if Tries .gt. Max\0E06then $ goto Wait_Too_Long X$ Tries =\0807+ 1 X$ wait 'Delay' X$ open/read/write/error=wait_loop tmp 'MXqueue' X$ close tmp X$ X$ X$! --------\0808\1010\2020\400C X$! Before we start the compress, we analyze the queue file to produce an X$! optimised version of the FDL. This is generated using the \3A06ze script X$! of the EDIT/FDL command and is used when regenerating the queue file X$! X$! NOTE: In case someone has grabbed access to the queue in the meantime, make V X$! a check and continue retrying up to the limit configured. X$! --------\0808\1010\2020\400C X$ em "-I-ANALYZE, ", f$fao("!8%T : Analyzing system queue FDL",0) X$ goto Anal_Begin X$Anal_Loop: X$ if Tries .gt. Max\0E06then $ goto Wait_Too_Long X$ Tries =\0807+ 1 X$ wait 'Delay' X$Anal_Begin: X$ analyze/rms/fdl/output='MXfdl'_temp 'MXqueue' X$ if $status .eq. RMS$_FLK then $ goto Anal_Loop X$ if .not. $status X$ then X$ em "-I-STATUS, ", f$fao("!8%T : !8XL : !AS", 0, $status, f$message( V\1307)) X$ em "-E-ANALFAIL, ", f$fao("!8%T : Analysis of system queue failed - abo Vrted",0) X$ delete 'MXfdl'_temp;* X$ goto CompressNotDone X$ endif X$ X$ X$ em "-I-OPTIMIZE, ", f$fao("!8%T : Optimizing the system queue FDL",0) X$ edit/fdl/nointeractive/analysis='MXfdl'_temp \0D0C X$ if .not. $status X$ then X$ em "-I-STATUS, ", f$fao("!8%T : !8XL : !AS", 0, $status, f$message( V\1307)) X$ em "-E-OPTIFAIL, ", f$fao("!8%T : Optimisation of system queue failed - V aborted",0) X$ delete 'MXfdl'_temp;* X$ goto CompressNotDone X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Now we can do the compress;\0A09 down to a temporary file first and X$! then rename it to the master one. This prevents unexpected aborts from X$! leaving partial queue files. X$! X$! NOTE: In case someone has grabbed access to the queue in the meantime, make V X$! a check and continue retrying up to the limit configured. X$! --------\0808\1010\2020\400C X$ em "-I-COMPRESS, ", f$fao("!8%T : Compressing MX queue file using optimised V FDL",0) X$ goto Compress_Begin X$Compress_Loop: X$ if Tries .gt. Max\0E06then $ goto Wait_Too_Long X$ Tries =\0807+ 1 X$ wait 'Delay' X$Compress_Begin: X$ convert/nosort - X /fdl='MXfdl'_temp - X`09'MXqueue' - X 'MXqueue'_temp X$ if $status .eq. RMS$_FLK then $ goto Compress_Loop X$ if .not. $status X$ then X$ em "-I-STATUS, ", f$fao("!8%T : !8XL : !AS", 0, $status, f$message( V\1307)) X$ em "-I-COMPFAIL, ", f$fao("!8%T : Unable to complete the queue compress V - aborted",0) X$ delete 'MXqueue'_temp;* X$ delete 'MXfdl'_temp;* X$ goto CompressNotDone X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Rename the newly built queue file to the master\1306and purge off old X$! copies X$! --------\0808\1010\2020\400C X$ em "-I-UPDATE, ", f$fao("!8%T : Updating queue file",0) X$ rename 'MXqueue'_temp\0F0A X$ purge 'MXqueue' X$ rename 'MXqueue' ;1 X$ X$ delete 'MXfdl'_temp;* X$ X$ X$! --------\0808\1010\2020\400C X$! Compress aborted, finished or timed out; restart MX X$! X$! The restart is complicated by two things. First, if the system is in a X$! cluster and the\1009mode is selected, it is necessary to restart MX X$! on all the nodes in the cluster that originally ran it. We do this by using V X$! either SYSMAN or a set of node specific BATCH queues. Second, this X$! procedure may be run by any user with the right privileges; however, MX mus Vt X$! be restarted under it's original username (see p4). If the requires\2309 X$! is the same as our current one, we use SYSMAN otherwise we submit the X$! startup procedure to each of the specified batch queues (see P1). X$! X$! --------\0808\1010\2020\400C X$CompressNotDone: X$Wait_Too_Long: X$RestartMx: X$ if f$edit(f$getjpi("","USERNAME"),"UPCASE,TRIM") .eqs. Username X$ then X$ X$ ! Issue the restart command on each node in the cluster using the sysman X$ ! command. Note that the MX_ROOT:MX_STARTUP_INFO.DAT file should X$ ! specifically name those nodes on which MX components should run because X$ ! this procedure has no knowledge of which nodes in the cluster run MX X$ ! and which don't. X$ em "-I-RESTART, ", f$fao("!8%T : Restarting MX on all nodes via SYSMAN V",0) X$ mc sysman set env/cluster X set timeout 00:00:30`09`09&0209! In case of hangs! X do @sys$startup:mx_\0B07`09`09`09! Restart cluster-wide X$ else X$ X$ ! Start up the restart job by submitting it to a batch queue appropriate to V X$ ! the node (see p1) X$ n=0 X$ QueueSubmit: X$ q = f$element(n, ",", Queues) X$ if q .nes. "," X$ then X$ x = f$getqui("CANCEL_OPERATION") X$ node = f$getqui("DISPLAY_QUEUE", "SCSNODE_NAME", q, "ALL_JOBS") X$ em "-I-RESTART, ", f$fao("!8%T : Restarting MX on node !AS via job V on batch queue !AS", 0, node, q ) X$ submit/user='Username'/nolog/queue='q' sys$startup:mx_\0B07 X$ n=n+1 X$ goto QueueSubmit X$ endif X$ endif X$ X$ X$! --------\0808\1010\2020\400C X$! Do a QUEUE PURGE to force out as many dead entries as possible. Note that X$! this should normally get done automatically by the ROUTER but sometimes it V X$! doesn't so this is a safety precaution. Note that it may take a LONG time o Vn X$! some systems and you may want to comment it out!!! X$! --------\0808\1010\2020\400C X$ em "-I-PURGING, ", f$fao("!8%T : Purging old entries",0) X$ mc mx_exe:mcp queue purge $ call unpack MX_SHRINK.COM;24 457217532 "" 23 1 2 $! $ create 'f' XMX QUEUE FILE COMPRESSION X X8-NOV-1993 X XAndy Harper`09`09&0209&0309Version 1.5 X X--------\0808\1010\2020\400F X X XINTRODUCTION X XThe queue file used by the MX mail system (MX_FLQ_DIR:SYSTEM_QUEUE.FLQ_CTL) Xwill, on a heavily used system, grow unbounded. This causes performance Xproblems due to excessive use of file extent pointers (hence lots of disk Xseeks). X XThe DCL procedure described here is an attempt to reduce the performance Xproblems by periodically compressing the queue file. X X X XOPERATION X XThe queue file cannot be compressed while MX is running (strictly,\1F07any Xuser has the file open) so MX needs to be shut down. The procedure will first V Xshut MX down, then compress the queue file, and finally restart MX. Depending V Xon the size of the queue file, MX may be unavailable for some considerable Xtime. It is recommended that this procedure be run at the quietest moment Xovernight. X XThe procedure attempts to be fail safe. When the shutdown signal has been sen Vt Xto MX, it loops trying to open the queue file. Once it can, no-one is accessin Vg Xit and this indicates that MX has completely shut down (after finishing Xwhatever it was doing in a tidy manner). However, should the procedure be Xunable to open the queue file after a number of retries, it will give up and Xrestart MX. X XMX is also restarted if the procedure is aborted or\1F08compression attempt Xfails in some way. X XOnce the queue is accessible to the procedure, the compression operation Xbegins. A new file is created using the currently valid records of the existin Vg Xqueue. This new file is given a different name (and later renamed) to avoid a V Xpartial queue file being inadvertently put in place as the master. X XThe OPTIMIZE script of EDIT/FDL is used to produce an optimal FDL file for the V Xsystem queue. This helps to ensure optimum performance. X X X X XRUNNING THE PROCEDURE X XThe procedure is run as follows: X X @MX_SHRINK `5Bbatch-queues`5D `5Bdelay`5D `5Bretry`5D `5Busername`5D X Xwhere: X X `5Bbatch-queues`5D X`09`09Specifies a comma separated list of batch queues into which X`09`09the MX restart job is submitted. Each queue should execute on X`09`09one of the nodes in the cluster and there should be one queue X`09`09for each node on which MX runs. X X`09`09If the node is stand-alone, then this paramter is OPTIONAL X`09`09and defaults to SYS$BATCH. In a cluster, a suitable list X`09`09of batch queues is required. These must be \2806execution X`09`09queues. X X`09`09Note this parameter is NOT used if the procedure is submitted X`09`09under the username given by the `5B\1708`5D parameter. Instead, X`09`09the procedure is submitted using SYSMAN. X X `5Bdelay`5D`09is the delta time to wait between retries `5Bdefault: 15 seco Vnds`5D X X `5Bretry`5D`09is the number of times to retry`09if the queue file cannot be V X`09`09opened (default: 240) X X `5Busername`5D`09is the name under which to restart MX `5Bdefault: SYSTEM V`5D X X XYou should arrange to run this procedure on a regular basis every night by som Ve Xexternal means. X X X XACKNOWLEDGEMENTS X XI'm grateful to Hunter Goatley, who provided the basic draft of this procedure V, Xfor permission to make his code freely available in my modified form. X XI'm grateful also to Dan WIng who has been dilligent in spotting typos and Xmaking useful suggestions. X`0C XCHANGES FROM VERSION 1.4 X X1.`09Fixed a type in a 'then" clause of an if. An erroneous "7" has X`09appeared. X X2.`09Do the purge of the queue file AFTER restarting MX rather than before X`09shutting it down. This allows the FDL optimisation to get a better idea X`09of a typical queue size. X X3.`09Fixed a typo that assigned the wrong parameter to the "Username" X`09symbol. X X4.`09Since adding the optimize fdl code, a number of timing windows were X`09introduced whereby a user could gain access to the MX queue file and X`09prevent successful operation. These windows have now been closed. X`09Note that this bug was non-fatal, it merely prevented completion of X`09the procedure. X X5.`09Added the current time to all messages. X X X XCHANGES FROM VERSION 1.3 X X1.`09Added automatic optimization of the queue file characteristics, using X`09the OPTIMIZE script of EDIT/FDL. X X X XCHANGES FROM VERSION 1.2 X X1.`09SETPRV privilege is needed to startup MX; we now check for this. X X2.`09Various typos fixed. X X3.`09Added /NOSORT to the CONVERT/FDL command to speed things up a bit X X4.`09Added a QUEUE PURGE command prior to shutting everything down. X X X XCHANGES FROM VERSION 1.1 X X1.`09SYSLCK privilege is required for an MX shutdown and SETPRV\2A0C X`09to restart MX; this is now checked before any actions are taken. X X2.`09The procedure now tries to deal with MX running in a cluster and using X`09a shared MX queue file by: X `09`09(a)`09Issuing a SHUTDOWN /CLUSTER command via MCP X`09`09(b)`09Restarting MX by submitting the MX_STARTUP job either X`09`09`09via SYSMAN or into a list of node specific batch queues X`09`09`09supplied as parameters. X X`09As a consequence, the parameter list of\1606rocedure has changed. X X`09See the documentation and source code for full details of how this X`09works. X X3.`09Fixed a typo in the documentation that transposed the parameters X`09`5Bretry`5D and `5Bdelay`5D. $ call unpack MX_SHRINK.DOC;10 627969271 "" 10 2 2 $ v=f$verify(v) $ exit ================================================================================ Archive-Date: Wed, 10 Nov 1993 12:22:59 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Tuning MX V3.3 queue file Message-ID: <1993Nov10.104531.317@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 10 Nov 93 10:45:31 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <1993Nov10.121349.18075@vax.oxford.ac.uk>, daveh@vax.oxford.ac.uk (Dave Hastings, OUCS) writes: >MX seems to work better if the queue file's cluster size is divisible by the >bucket size (On our system the cluster size is 51 blocks and the bucket size is >17 blocks). Are you saying your *disk* cluster size is 51 blocks? >> 4k. The FDL editor will then spit out an informational message >> about the depth of key 0, and other information. Press RETURN >> to continue. >> <RETURN> > >The index should be as flat as possible. I agree; I'll make the necessary changes to the text so that is done. ... I'm also adding instructions to this document on why you'd want to run multiple MX agents (SMTP, Router, Local, etc.) to increase throughput for higher-volume systems. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Thu, 11 Nov 1993 06:55:49 CST Sender: list-mgr@WKUVX1.WKU.EDU Subject: RE: MX queue not auto-purging Message-ID: <1993Nov11.102703.18107@vax.oxford.ac.uk> From: daveh@vax.oxford.ac.uk (Dave Hastings, OUCS) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 11 Nov 93 10:27:03 GMT To: MX-List@WKUVX1.WKU.EDU In article <17327811@MVB.SAIC.COM>, "Brian Tillman" <tillman@swdev.si.com> writes: >>"Andy, Systems Manager" <udaa055@BAY.CC.KCL.AC.UK> writes: >>> >>>I seem to have a situation whereby my MX queue file is not automatically >>>clearing out the old FINISHED entries! As a result, my queue file grows quite >>>large! I do not have the appropriate MX logical defined and, according to the >>>documentation, it should default to clearing out every 15 minutes. >>> >>>Could anyone suggest what the problem might be? At the moment I'm having to do >>>a manual QUEUE PURGE and it can take a long long time if you don't do it >>>frequently! >>> >>I had this same thing happen to me today. The problem here, and most >>likely yours, is that there are so many messages to be processed that >>the Router never gets finished with new messages, so it never gets >>around to the PURGE. > > This very thing also happened to me yesterday. A _ton_ of messages came in and, > even though I run two routers (one on each of two machines in a cluster), one > router was doing nothing but purging finished entries. I had to wait about 40 > minutes after a SHUTDOWN for one of the routers to finish what it was doing. > Trying to use MCP QUEUE SHOW during this time was a total waste of time. > -- > Brian Tillman > tillman_brian@si.com > We hit this problem about a year ago, and got round it by running a separate detached process which does a MCP QUEUE PURGE more or less constantly. This takes a load off the Router. You have to modify the startup code to start this process, and you have to explicitly close it down if you shut down MX. This is the code that we use. The process name is MX_Purge'n', where 'n' is incremented by one every time the loop is executed. Usually each purge pass takes about 2 minutes. You need to increase the value of the MX_FLQ_PURGE_WAIT logical, so that the Router doesn't sneak in a purge while the MX_Purge process is waiting. Dave -- $! command procedure to purge the MX queue. Run as a detached process $! $ set nover $ on error then goto ERR $ assign/job "sys$sysroot:[sysmgr]" sys$login $ assign/job "sys$sysroot" sys$login_device $ assign/job "sys$sysroot:[sysmgr]" sys$scratch $ mcp :==$mx_exe:mcp $ count=0 $loop: $ set proc/name="MX_Purge''count'" $ sh time $ write sys$output "Purging..." $ mcp queue purge $ sh time $ count=count+1 $ if f$trnlnm("mx_oucs_purge_wait") .eqs. "" $ then $ wait_time="00:00:45" $ else $ wait_time=f$trnlnm("mx_oucs_purge_wait") $ endif $ write sys$output "Waiting: ''wait_time'" $ wait 'wait_time' $ goto LOOP $ exit $ERR: $ @sys$sysroot:[oxford]start_mx_queue_purge $ exit 42 -- David Hastings | "There's nothing wrong with my sense VAX Systems Programmer | of reality - I have it thoroughly Oxford University Computing Services| serviced once a fortnight" - DNA daveh@vax.oxford.ac.uk | ================================================================================ Archive-Date: Thu, 11 Nov 1993 07:38:02 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: MX_SHRINK 1.5 comment Message-ID: <CGBuAs.245@curia.ucc.ie> From: chris@csvax1.ucc.ie (Chris Higgins - System Administrator) Date: Thu, 11 Nov 1993 12:20:52 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU "Andy, Systems Manager" <udaa055@bay.cc.kcl.ac.uk> writes: >Dear All, > >Thank you for your helpful comments on the MX_SHRINK procedure recently posted. >Enclosed is the latest version which fixes a few typos and sorts out a minor >timing window. > >If no bugs are found, I plan to make this the last version. One small point. I've added " $set symbol/scope=(nolocal,noglobal) " to the start of MX_SHRINK.COM, because I have EDIT defined to be EDIT/TPU and this caused the EDIT/FDL line to bomb out.... > >Regards, > >Andy Harper >Kings College London Chris. + J.C. Higgins, + Chris@cs.ucc.ie + If you love something, set it + + VMS Sys. Admin, + Chris@csvax1.ucc.ie + free. If it doesn't come back + + Comp.Sc.Dept. + Chris@odyssey.ucc.ie + to you, hunt it down and + + UCC, Ireland + C.Higgins@bureau.ucc.ie + KILL it. + ================================================================================ Archive-Date: Thu, 11 Nov 1993 07:38:08 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: RE: MX queue not auto-purging Message-ID: <1993Nov11.114443.18110@vax.oxford.ac.uk> From: daveh@vax.oxford.ac.uk (Dave Hastings, OUCS) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 11 Nov 93 11:44:43 GMT To: MX-List@WKUVX1.WKU.EDU In article <1993Nov11.102703.18107@vax.oxford.ac.uk>, daveh@vax.oxford.ac.uk (Dave Hastings, OUCS) writes: > In article <17327811@MVB.SAIC.COM>, "Brian Tillman" <tillman@swdev.si.com> writes: >>>"Andy, Systems Manager" <udaa055@BAY.CC.KCL.AC.UK> writes: >>>> >>>>I seem to have a situation whereby my MX queue file is not automatically >>>>clearing out the old FINISHED entries! As a result, my queue file grows quite >>>>large! I do not have the appropriate MX logical defined and, according to the >>>>documentation, it should default to clearing out every 15 minutes. >>>> >>>>Could anyone suggest what the problem might be? At the moment I'm having to do >>>>a manual QUEUE PURGE and it can take a long long time if you don't do it >>>>frequently! >>>> >>>I had this same thing happen to me today. The problem here, and most >>>likely yours, is that there are so many messages to be processed that >>>the Router never gets finished with new messages, so it never gets >>>around to the PURGE. >> >> This very thing also happened to me yesterday. A _ton_ of messages came in and, >> even though I run two routers (one on each of two machines in a cluster), one >> router was doing nothing but purging finished entries. I had to wait about 40 >> minutes after a SHUTDOWN for one of the routers to finish what it was doing. >> Trying to use MCP QUEUE SHOW during this time was a total waste of time. >> -- >> Brian Tillman >> tillman_brian@si.com >> > > We hit this problem about a year ago, and got round it by running a separate > detached process which does a MCP QUEUE PURGE more or less constantly. This > takes a load off the Router. You have to modify the startup code to start this > process, and you have to explicitly close it down if you shut down MX. > > This is the code that we use. The process name is MX_Purge'n', where 'n' is > incremented by one every time the loop is executed. > > Usually each purge pass takes about 2 minutes. > > You need to increase the value of the MX_FLQ_PURGE_WAIT logical, so that the > Router doesn't sneak in a purge while the MX_Purge process is waiting. Oops! The MX_FLQ_CHECK_WAIT logical needs to be high (eg 1 day) and the MX_FLQ_PURGE_WAIT logical needs to be low (eg 10 seconds). Dave -- David Hastings | "There's nothing wrong with my sense VAX Systems Programmer | of reality - I have it thoroughly Oxford University Computing Services| serviced once a fortnight" - DNA daveh@vax.oxford.ac.uk | ================================================================================ Archive-Date: Thu, 11 Nov 1993 12:10:29 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Thu, 11 Nov 1993 13:05:20 EDT From: "Wheeler, Daniel A" <manager@LEMOYNE.BITNET> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-LIST@WKUVX1.WKU.EDU CC: manager@MAPLE Message-ID: <0097561B.DFFCC720.19438@lemoyne> Subject: MX Configuration - Adding Internet? OK, I'm lost again and would appreciate a steer to an answer and the references for proper MX configuration. I'm in the process of adding direct Internet service to our continuing BITNET service on MX 3.3/VMS 5.5-2 . We are LEMOYNE.BITNET, maple.lemoyne.edu, and lemoyne.edu. My mail seems to be sent with a From: <user@maple> and a Sender: user@maple.lemoyne.edu. The From address results in remote systems attempting to reply to a local node maple. Direct addressing from the remote system works fine. I can't even figure why the From field contains just maple. Is there someplace in my config that I have set this? MX configs and logical names are listed below. ---dan --------------------------------------------------------------------------- Dan Wheeler Senior Support Analyst & VMS System Manager Office of Information Systems, Le Moyne College, Syracuse NY 13214-1399 Phone 315.445.4565 (Home 315.655.8193) wheeler@maple.lemoyne.edu --------------------------------------------------------------------------- =========================================================================== Configuration file: MX_DEVICE:[MX]MX_CONFIG.MXCFG;44 MX version id is: MX V3.3 VAX Mailing lists: ... Address-rewriting rules: Rewrite "<{u}@lemoyne>" => "<{u}@lemoyne>" Rewrite "<{u}@lemoyne.BITNET>" => "<{u}@lemoyne>" Rewrite "<{u}@maple.lemoyne.edu>" => "<{u}@lemoyne>" Rewrite "<{folder}@BULLETIN>" => "<{folder}@BULLETIN>" Rewrite "<@{stuff}:{user}@{host}.UUCP>" => "<{user}%{host}%uunet.uu.net@INTERBIT.BITNET>" Rewrite "<{user}@{host}.UUCP>" => "<{user}%{host}%uunet.uu.net@INTERBIT.BITNET>" Rewrite "<@{stuff}:{user}@{host}.CSNET>" => "<{user}%{host}.CSNET%relay.cs.net@INTERBIT.BITNET>" Rewrite "<{user}@{host}.CSNET>" => "<{user}%{host}.CSNET%relay.cs.net@INTERBIT.BITNET>" Rewrite "<@{route}:{user}@{host}.BITNET>" => "<{user}@{host}.BITNET>" Rewrite "<{user}@{host}.BITNET>" => "<{user}@{host}.BITNET>" Rewrite "<@{route}:{user}@{host}>" => "<{user}%{host}@INTERBIT.BITNET>" Rewrite "<{U}@SOCRATES>" => "<{U}@LEMOYNE>" Rewrite "<{U}@MAPLE>" => "<{U}@LEMOYNE>" Rewrite "<{U}@admin1>" => "<{U}@oak>" Domain-to-path mappings: Domain="[192.231.122.100]", Path=Local Domain="maple", Path=Local Domain="lemoyne", Path=Local Domain="lemoyne.edu", Path=Local Domain="maple.lemoyne.edu", Path=Local Domain="BULLETIN", Path=Site Domain="oak", Path=SMTP Domain="oak.lemoyne.edu", Path=SMTP Domain="lemoyne.bitnet", Path=Local Domain="socrates", Path=Local Domain="admin1", Path=SMTP Domain="*.edu", Path=SMTP Domain="*.uucp", Path=SMTP, Route="uunet.uu.net" Domain="*.bitnet", Path=Jnet Domain="lmccsc", Path=Local Domain="vax.cs.hscsyr.edu", Path=SMTP Domain="*.syr.edu", Path=SMTP Domain="*.psi.com", Path=SMTP Domain="BAMBOO", Path=SMTP Domain="bamboo.lemoyne.edu", Path=SMTP Domain="lehigh.edu", Path=SMTP Domain="nysernet.org", Path=SMTP SMTP agent settings: Retry interval: 0 00:15:00.00 Maximum number of retries: 96 Number of DNS failure retries: 12 Accounting: enabled Default router: (none) LOCAL agent settings: DECnet delivery retry interval: 0 00:05:00.00 Maximum number of retries: 96 Accounting enabled. Multiple VMS Mail From: addresses allowed. Top headers: FROM,SENDER,TO,RESENT_TO,CC,RESENT_CC,BCC,RESENT_BCC,MESSAGE_ID, RESENT_MESSAGE_ID,IN_REPLY_TO,REFERENCES,KEYWORDS,SUBJECT, ENCRYPTED,DATE,REPLY_TO,RECEIVED,RESENT_REPLY_TO,RESENT_FROM, RESENT_SENDER,RESENT_DATE,RETURN_PATH,OTHER Bottom headers: (none) ROUTER agent settings: Automatic percent-hack handling: enabled JNET agent settings: Automatic percent-hack handling: disabled BSMTP replies: disabled Accounting: enabled Lenient about gatewaying mail: yes Primary mailer username: MXMAILER DECnet_SMTP agent settings: Retry interval: 0 00:30:00.00 Maximum number of retries: 96 Accounting disabled. SITE agent settings: Retry interval: 0 00:01:00.00 Maximum number of retries: 96 X25_SMTP agent settings: Retry interval: 0 00:30:00.00 Maximum number of retries: 96 Accounting disabled. =========================================================================== MX_FLQ_NODE_NAME\/SYSTEM/EXEC\NEWKID MX_NODE_NAME\/SYSTEM/EXEC\lemoyne MX_VMSMAIL_LOCALHOST\/SYSTEM/EXEC\@maple.lemoyne.edu MX_FLQ_DIR\/SYSTEM/EXEC\DISK$APPS:[MX.QUEUE] MX_FLQ_RECLAIM_WAIT\/SYSTEM/EXEC\0 00:15:00 MX_ROUTER_DIR\/SYSTEM/EXEC\MX_ROOT:[ROUTER] MX_LOCAL_DIR\/SYSTEM/EXEC\MX_ROOT:[LOCAL] MX_MCP_HELPLIB\/SYSTEM\MX_DIR:MX_MCP_HELPLIB MX_ALIAS_HELPLIB\/SYSTEM\MX_DIR:MX_ALIAS_HELPLIB =========================================================================== (LNM$SYSTEM_TABLE) [kernel] [shareable,system] [Protection=(RWE,RWE,R,R)] [Owner=[ISS,SYSTEM]] "MX_ALIAS_HELPLIB" [super] = "MX_DIR:MX_ALIAS_HELPLIB" "MX_DEVICE" [exec] = "NEWKID$DKA200:" [concealed,terminal] "MX_DIR" [exec] = "MX_DEVICE:[MX]" "MX_DOC" [exec] = "MX_ROOT:[DOC]" "MX_EXAMPLES_DIR" [exec] = "MX_ROOT:[EXAMPLES]" "MX_EXE" [exec] = "MX_ROOT:[EXE]" "MX_FLQ_DIR" [exec] = "DISK$APPS:[MX.QUEUE]" "MX_FLQ_NODE_NAME" [exec] = "NEWKID" "MX_FLQ_RECLAIM_WAIT" [exec] = "0 00:15:00" "MX_FLQ_SHR" [exec] = "MX_EXE:MX_FLQ_SHR" "MX_JNET_DIR" [exec] = "MX_ROOT:[JNET]" "MX_LOCAL_DEBUG" [exec] = "1" "MX_LOCAL_DIR" [exec] = "MX_ROOT:[LOCAL]" "MX_MAILSHR" [exec] = "MX_EXE:MX_MAILSHR" "MX_MAILSHRP" [exec] = "MX_EXE:MX_MAILSHRP" "MX_MCP_HELPLIB" [super] = "MX_DIR:MX_MCP_HELPLIB" "MX_MLF_DIR" [exec] = "MX_ROOT:[MLF]" "MX_MLIST_DIR" [exec] = "MX_ROOT:[MLF.MAILING_LISTS]" "MX_MSG" [exec] = "MX_EXE:MX_MSG" "MX_NODE_NAME" [exec] = "lemoyne" "MX_ROOT" [exec] = "MX_DEVICE:[MX.]" [concealed] "MX_ROUTER_DEBUG" [exec] = "1" "MX_ROUTER_DIR" [exec] = "MX_ROOT:[ROUTER]" "MX_SHR" [exec] = "MX_EXE:MX_SHR" "MX_SITE_DEBUG" [exec] = "1" "MX_SITE_DIR" [exec] = "MX_ROOT:[SITE]" "MX_SITE_DOM_EXPANSION" [exec] = "MX_EXE:DOMAIN_EXPANSION" "MX_SMTP_DEBUG" [exec] = "1" "MX_SMTP_DIR" [exec] = "MX_ROOT:[SMTP]" "MX_UUCP_DIR" [exec] = "MX_ROOT:[UUCP]" "MX_VMSMAIL_LOCALHOST" [exec] = "@maple.lemoyne.edu" ================================================================================ Archive-Date: Thu, 11 Nov 1993 12:23:53 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Thu, 11 Nov 1993 12:23:08 CST From: "Hunter Goatley, WKU" <goathunter@WKUVX1.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: manager@maple.lemoyne.edu Message-ID: <00975615.FB070E6E.19254@WKUVX1.WKU.EDU> Subject: RE: MX Configuration - Adding Internet? "Wheeler, Daniel A" <manager@LEMOYNE.BITNET> writes: > >OK, I'm lost again and would appreciate a steer to an answer and the >references for proper MX configuration. > >I'm in the process of adding direct Internet service to our continuing >BITNET service on MX 3.3/VMS 5.5-2 . We are LEMOYNE.BITNET, >maple.lemoyne.edu, and lemoyne.edu. My mail seems to be sent with a >From: <user@maple> and a >Sender: user@maple.lemoyne.edu. > >The From address results in remote systems attempting to reply to a >local node maple. Direct addressing from the remote system works fine. > Are you using UCX? I'd guess that you are, and that you have the the node name MAPLE listed first in the local UCX host table. This is described in the MX release notes; you need to remove the name MAPLE and then add the FQDN (maple.lemoyne.edu) as the first name. Extract included below. >MX_NODE_NAME\/SYSTEM/EXEC\lemoyne Also, this should be maple.lemoybe.edu. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ------------------------------------------------------------------------------- 1.3 VMS/ULTRIX Connection Notes If you are using DEC VMS/ULTRIX Connection (UCX) V1.2 through V2.0, you should review your UCX configuration to ensure that you are using your fully-qualified domain name (FQDN) as your primary local host name in your UCX host table. If you followed the UCX documentation when configuring UCX with UCX$CONFIG, it is highly likely that you did not specify your FQDN as your primary host name. If your abbreviated host name is listed as the primary in your UCX host table, then messages sent from the local system will go out with an unreplyable return address. To check your local host name, use the UCX utility: $ UCX UCX> SHOW HOST/LOCAL x.x.x.x ! use your IP address here To ensure that your FQDN is the primary host name for your address, use the UCX commands: UCX> SET NOHOST x.x.x.x ! use your IP address here UCX> SET HOST "full.qual.dom.name"/ADDR=x.x.x.x/ALIAS="abbrev" You will be asked for confirmation on the SET NOHOST command. Substitute your FQDN and IP address in the SET HOST command, and if you would still have your shortened host name in the host table, use the /ALIAS qualifier, as shown. Once you have updated your host table, you should review the file SYS$MANAGER:UCX$INET_SET_ INTERFACES.COM and replace all instances of your abbreviated host name with your FQDN. Note: Remember that UCX is case sensitive with regard to host names! ================================================================================ Archive-Date: Fri, 12 Nov 1993 02:13:13 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: MX Configuration - Adding Internet? Message-ID: <2bva32$k9t@gap.cco.caltech.edu> From: carl@SOL1.GPS.CALTECH.EDU (Carl J Lydick) Date: 12 Nov 1993 06:22:26 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17388562@MVB.SAIC.COM>, "Wheeler, Daniel A" <manager@LEMOYNE.BITNET> writes: =OK, I'm lost again and would appreciate a steer to an answer and the =references for proper MX configuration. = =I'm in the process of adding direct Internet service to our continuing =BITNET service on MX 3.3/VMS 5.5-2 . We are LEMOYNE.BITNET, =maple.lemoyne.edu, and lemoyne.edu. My mail seems to be sent with a =From: <user@maple> and a =Sender: user@maple.lemoyne.edu. = =The From address results in remote systems attempting to reply to a =local node maple. Direct addressing from the remote system works fine. = =I can't even figure why the From field contains just maple. Is there =someplace in my config that I have set this? MX configs and logical =names are listed below. You don't tell us what your TCP/IP package is. At a guess, it's UCX. If so, look at the MX release notes for instructions on the surgery you need to perform as a result of having followed DEC's instructions on how to shoot yourself in the foot when setting up UCX. -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Fri, 12 Nov 1993 06:50:35 CDT Sender: list-mgr@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU From: "Eric R. Rountree" <ROUNTREE@SBACOOP.SBA.DAL.CA> Date: 12 Nov 1993 8:49:17 AST Subject: Messages stuck in queue Reply-To: MX-List@WKUVX1.WKU.EDU Message-ID: <D8C3FB06F6@SBACOOP.SBA.DAL.CA> Does anyone know what is happening when a message is stuck in MX's queue with the error: "Message number 00008100" attached to it? Thanks Eric +-----------------------------------------+ | Eric R. Rountree, System Manager | | | | School of Business Administration Site | | Distributed Sites Unit | | Academic Computing Services | | Dalhousie University | | Halifax, Nova Scotia | | | | Rountree@SBACOOP.SBA.Dal.CA | +-----------------------------------------+ ================================================================================ Archive-Date: Fri, 12 Nov 1993 09:42:00 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Messages stuck in queue Message-ID: <1993Nov12.080025.340@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 12 Nov 93 08:00:25 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17416568@MVB.SAIC.COM>, "Eric R. Rountree" <ROUNTREE@SBACOOP.SBA.DAL.CA> writes: >Does anyone know what is happening when a message is stuck in MX's queue >with the error: "Message number 00008100" attached to it? TCPware shows this error as: $ SET MESSAGE SYS$MESSAGE:TCPWARE_MSG $ EXIT %X08100 %TCPWARE-W-EPIPE, pipe is broken -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Fri, 12 Nov 1993 14:28:34 CDT Sender: list-mgr@WKUVX1.WKU.EDU Message-ID: <9311122022.AA02010@snoopy.ucis.dal.ca> To: MX-List@WKUVX1.WKU.EDU From: "Eric Rountree" <ROUNTREE@SBACOOP.SBA.DAL.CA> Reply-To: MX-List@WKUVX1.WKU.EDU Date: Fri, 12 Nov 1993 16:22:30 AST Subject: Re: Messages stuck in queue I wrote: > Does anyone know what is happening when a message is stuck in > MX's queue with the error: "Message number 00008100" attached to > it? Dan Wing replied: > TCPware shows this error as: > > $ SET MESSAGE SYS$MESSAGE:TCPWARE_MSG > $ EXIT %X08100 > %TCPWARE-W-EPIPE, pipe is broken Hmm. I'm not running TCPware. I'm running Multinet. Does anyone know what I should look for here? I'll provide as much information as I can, but I'm not sure what information I need to provide. Thanks Eric +-----------------------------------------+ | Eric R. Rountree, System Manager | | | | School of Business Administration Site | | Distributed Sites Unit | | Academic Computing Services | | Dalhousie University | | Halifax, Nova Scotia | | | | Rountree@SBACOOP.SBA.Dal.CA | +-----------------------------------------+ ================================================================================ Archive-Date: Fri, 12 Nov 1993 16:20:13 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Messages stuck in queue Message-ID: <1993Nov12.145746.349@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 12 Nov 93 14:57:45 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17425630@MVB.SAIC.COM>, "Eric Rountree" <ROUNTREE@SBACOOP.SBA.DAL.CA> writes: >I wrote: >> Does anyone know what is happening when a message is stuck in >> MX's queue with the error: "Message number 00008100" attached to >> it? > >Dan Wing replied: >> TCPware shows this error as: >> >> $ SET MESSAGE SYS$MESSAGE:TCPWARE_MSG >> $ EXIT %X08100 >> %TCPWARE-W-EPIPE, pipe is broken > >Hmm. I'm not running TCPware. I'm running Multinet. Does anyone >know what I should look for here? I'll provide as much information as I >can, but I'm not sure what information I need to provide. Does this happen for all messages? Enable debugging for all MX agents, and see which agent is returning the error code. The MX Management Guide describes how to enable debugging. After you get some log files generated, you can do: $ SET DEFAULT MX_ROOT:[000000] $ SEARCH [.*]*.LOG;* 08100 and post your results. (And, so we don't bounce posts back and forth more, what version of MX, VMS, and MultiNet are you using??) -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Mon, 15 Nov 1993 12:35:51 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 15 Nov 1993 13:31:14 -0500 (EST) From: Teresa Feck <D2DCFEC@FRE.TOWSON.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU Subject: Efficiency problem with mail queue To: mx-list@wkuvx1.wku.edu Message-ID: <01H5CGZUSV8I8WYAU0@TOE.TOWSON.EDU> MIME-Version: 1.0 Content-Transfer-Encoding: 7BIT Date sent: 15-NOV-1993 13:28:17 I am running MX 3.3 on a VAX 4300. Our system is running under VMS 5.5-2 and UCX 2.0. This system is not a cluster system and were are not running Jnet. We are experiencing a major delay with mx mail processing through the queue. I have installed various listsevers on the system. Most of the listservers are fairly small but there is one that contains 460 subscribers which are all nonlocal subscribers. This list seems to have to the most problems with delays. The delays have been as large as 2 weeks from the time someone sends something to the list to the time it is received. I put into place the compression command file that was published on the list server weeks ago and I have it executing daily. It seems to have helped a little but we are still having problems with delays. Eventhough, our largest list is the one that is encountering most of the problems, it takes small messages several hours to process. Also, to assist in speeding the mail queue I have modified the cleanup logicals MX_FLQ_PURGE_WAIT to 5 minutes and MX_FLQ_RECLAIM_WAIT to 30 minutes. Located below is a listing of my configuration files. I have excluded the listing of all listservers with the exception of the one that is encountering the largest delays (prolist). As you can see, I have modified the DNS and error count retries. I hope I've furnished enough information that will allow someone to assist me. Would it help to increase the MX_SMTP_SERVER_THREADS or to run mulitple SMTP agents. Any help would be greatly appreciated!!!! Configuration file: MX_DEVICE:[MX]MX_CONFIG.MXCFG;74 MX version id is: MX V3.3 VAX Mailing lists: Name: prolist Owner: "prolistowner@FRE.FSU.UMD.EDU" Reply-to: List, NOSender Archive: DISK02:[000000.PROLIST] Add message: MX_MLIST_DIR:PROLIST_ADD.TXT Remove message: MX_MLIST_DIR:PROLIST_REMOVE.TXT Errors-to: PROLISTOWNER@FRE.FSU.UMD.EDU Strip header: NOReceived Private list: No Protection: (SYSTEM:RWED,OWNER:RWED,GROUP:RWED,WORLD:WE) Address-rewriting rules: Rewrite "<@{route}:{user}@{host}.UUCP>" => "<@{route}:{user}@{host}.UUCP>" Rewrite "<{user}@{host}.UUCP>" => "<@uunet.uu.net:{user}@{host}>" Rewrite "<@{route}:{user}@{host}.CSNET>" => "<@{route}:{user}@{host}.CSNET>" Rewrite "<{user}@{host}.CSNET>" => "<@relay.cs.net:{user}@{host}.CSNET>" Rewrite "<@{route}:{user}@{host}.BITNET>" => "<@{route}:{user}@{host}.BITNET>" Rewrite "<{user}@{host}.BITNET>" => "<@cunyvm.cuny.edu:{user}@{host}.BITNET>" Rewrite "<{u}@DIGEST.SITE>" => "<{u}@DIGEST.SITE>" Domain-to-path mappings: Domain="fre.fsu.umd.edu", Path=Local Domain="fre", Path=Local Domain="fre00", Path=Local Domain="[131.118.80.2]", Path=Local Domain="*", Path=SMTP Domain="*.BITNET", Path=SMTP, Route="CUNYVM.CUNY.EDU" Domain="*.UUCP", Path=SMTP, Route="UUNET.UU.NET" Domain="DIGEST.SITE", Path=Site, Route="DIGEST" Aliases: LocalName="POSTMASTER", Address="d2dcfec@fre.fsu.umd.edu" File servers: Name: PROLIST_FILES, Manager: prolist@FRE.FSU.UMD.EDU Root: DISK02:[000000.PROLIST_FILES.] Not linked to a mailing list. Delay threshold: 0 Send period: 00:00 - 23:00 Daily limits: Server: 0 Host: 0 User: 0 SMTP agent settings: Retry interval: 0 00:30:00.00 Maximum number of retries: 5 Number of DNS failure retries: 4 Accounting: enabled Default router: (none) LOCAL agent settings: DECnet delivery retry interval: 0 00:30:00.00 Maximum number of retries: 96 Accounting enabled. Multiple VMS Mail From: addresses allowed. Top headers: FROM,SENDER,TO,RESENT_TO,CC,RESENT_CC,BCC,RESENT_BCC,MESSAGE_ID, RESENT_MESSAGE_ID,IN_REPLY_TO,REFERENCES,KEYWORDS,SUBJECT, ENCRYPTED,DATE,REPLY_TO,RECEIVED,RESENT_REPLY_TO,RESENT_FROM, RESENT_SENDER,RESENT_DATE,RETURN_PATH,OTHER Bottom headers: (none) ROUTER agent settings: Automatic percent-hack handling: enabled JNET agent settings: Automatic percent-hack handling: enabled BSMTP replies: disabled Accounting: disabled Lenient about gatewaying mail: no No mailer username set. DECnet_SMTP agent settings: Retry interval: 0 00:30:00.00 Maximum number of retries: 96 Accounting disabled. SITE agent settings: Retry interval: 0 00:30:00.00 Maximum number of retries: 5 X25_SMTP agent settings: Retry interval: 0 00:30:00.00 Maximum number of retries: 96 Accounting disabled. _______________________________ Teresa J. Feck Internet: feck@fre.fsu.umd.edu Academic Computing Bitnet: feck@fre.towson.edu Frostburg State University Phonemail: (301) 689-7090 Frostburg, Maryland 21532 ================================================================================ Archive-Date: Mon, 15 Nov 1993 13:52:05 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Efficiency problem with mail queue Message-ID: <1993Nov15.122019.361@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 15 Nov 93 12:20:19 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17534891@MVB.SAIC.COM>, Teresa Feck <D2DCFEC@FRE.TOWSON.EDU> writes: >Date sent: 15-NOV-1993 13:28:17 > I am running MX 3.3 on a VAX 4300. Our system is running under VMS > 5.5-2 and UCX 2.0. This system is not a cluster system and were are not >running Jnet. We are experiencing a major delay with mx mail processing >through the queue. How many outgoing SMTP agents have you configured? I'd recommend setting up at *least* two for any MX configuration, and yours might benefit from more -- if you have two, one of them can be busy sending a message to all your 460 subscribers, and the other can be sending out other things at the same time. I have a prelimiary copy of my "Tuning MX" document which I'll send you via mail (it includes information on setting up multiple MX agents, and why you'd want to setup multiple agents). -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Mon, 15 Nov 1993 14:17:01 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 15 Nov 1993 14:14:43 -0600 (CST) From: "H.A. Kippenhan Jr." <KIPPENHAN@fndcd.fnal.gov> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <931115141443.2a80446a@fndcd.fnal.gov> Subject: Re: Efficiency problem with mail queue Hi Dan: | | From: SMTP%"MX-List@WKUVX1.WKU.EDU" 15-NOV-1993 14:09:26.40 | Subj: Re: Efficiency problem with mail queue | | In article <17534891@MVB.SAIC.COM>, Teresa Feck <D2DCFEC@FRE.TOWSON.EDU> writes: | | >Date sent: 15-NOV-1993 13:28:17 | > I am running MX 3.3 on a VAX 4300. Our system is running under VMS | > 5.5-2 and UCX 2.0. This system is not a cluster system and were are not | >running Jnet. We are experiencing a major delay with mx mail processing | >through the queue. | | How many outgoing SMTP agents have you configured? I'd recommend setting | up at *least* two for any MX configuration, and yours might benefit from | more -- if you have two, one of them can be busy sending a message to all | your 460 subscribers, and the other can be sending out other things at | the same time. | | I have a preliminary copy of my "Tuning MX" document which I'll send you via | mail (it includes information on setting up multiple MX agents, and why | you'd want to setup multiple agents). | Can I get a copy as well. In return, I'll proof it and feed back my comments (if you would like). Regards - Kipp - {~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~} { H.A. Kippenhan Jr. | Internet: Kippenhan@FNDCD.FNAL.GOV } { National HEPnet Management | HEPnet/NSI DECnet: FNDCD::KIPPENHAN } { Fermi National Accelerator Lab. | BITnet: Kippenhan@FNDCD.BITNET } { P.O. Box 500 MS: FCC-3E/368 | Telephone: (708) 840-8068 } { Batavia, Illinois 60510 | FAX: (708) 840-8463 } ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ================================================================================ Archive-Date: Mon, 15 Nov 1993 14:41:51 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 15 Nov 1993 14:35:21 -0600 (CST) From: "The Eyes of Texas: (.)(.)" <stewart.nichols@dir.texas.gov> Reply-To: MX-List@WKUVX1.WKU.EDU Subject: RE: Efficiency problem with mail queue To: MX-List@WKUVX1.WKU.EDU CC: "The Eyes of Texas: (.)(.)" <NICHOLS_SA@DIR.TEXAS.GOV> Message-ID: <01H5CJBF4DXGB8JK81@DIR.TEXAS.GOV> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII Content-Transfer-Encoding: 7BIT > I have a prelimiary copy of my "Tuning MX" document which I'll send you via ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > mail (it includes information on setting up multiple MX agents, and why > you'd want to setup multiple agents). Could I have a copy of this, too? stu *--------------------------------*-------------------------------------* * Stewart Nichols * * * Wide Area Network Specialist * * * Dept. of Info Resources, Texas * * * Work Phone: 512-463-7601 * * * stewart.nichols@dir.texas.gov * * *--------------------------------*-------------------------------------* ================================================================================ Archive-Date: Mon, 15 Nov 1993 14:42:42 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 15 Nov 1993 15:48:27 -0600 To: mx-list@wkuvx1.wku.edu From: reef@xenon.che.ilstu.edu Reply-To: MX-List@WKUVX1.WKU.EDU Subject: Re: Efficiency problem with mail queue >I have a prelimiary copy of my "Tuning MX" document which I'll send you via >mail (it includes information on setting up multiple MX agents, and why >you'd want to setup multiple agents). > >-Dan Wing, Systems Administrator, University Hospital, Denver > dwing@uh01.colorado.edu or wing@eisner.decus.org > > Vote to sacrifice Ratbert! Send Email to scottadams@aol.com Send me one too!! Thanks, Reef *************************************************************************** Philip D. (Reef) Morse II Internet address: reef@xenon.che.ilstu.edu Department of Chemistry Telephone (voice and voice mail): (309) 438-5595 Illinois State University Telephone (FAX): (309) 438-5538 Normal, IL 61761 Telephone (answering machine): (309) 829-9257 USA *************************************************************************** ================================================================================ Archive-Date: Mon, 15 Nov 1993 15:45:06 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 15 Nov 1993 16:42:35 EST From: "Steve Thompson, Cheme System Mangler" <olin@cheme.cornell.edu> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: olin@cheme.cornell.edu Message-ID: <0097595E.E33BDFD4.20894@cheme.cornell.edu> Subject: RE: Efficiency problem with mail queue Dan sez: > I have a prelimiary copy of my "Tuning MX" document which I'll send you via > mail (it includes information on setting up multiple MX agents, and why > you'd want to setup multiple agents). Looks like you gotta post it to the list, Dan! -steve ================================================================================ Archive-Date: Mon, 15 Nov 1993 16:42:54 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: MX Tuning document (draft) Message-ID: <1993Nov15.150754.364@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 15 Nov 93 15:07:54 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU I'm sending the draft copy of the MX Tuning document to everyone that has asked. If you receive a copy, please make *some* sortof comment about it -- hard to read, difficult to understand, didn't work with your MX configuration, didn't like the author's last name, don't know who Ratbert is, etc. If you have other ideas for tuning MX that you think should be incorporated, let me know and I'll add 'em (with attributes to the author, of course). Right now it only contains two major sections: 1. Multiple MX agents - why you want them on your system 2. Tuning the MX queue file I'm considering adding some stuff about using RMS Global buffers, which has the potential to help a lot if all your MX agents are on the same VAXcluster node. If you want a copy, be sure to Email me *directly* (not the list!). Thanks. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Mon, 15 Nov 1993 18:09:01 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 15 Nov 1993 23:54:37 GMT From: Jamie Jones <jamie@hicom.lut.ac.uk> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU Message-ID: <0097599B.3DFCF520.1640@hicom.lut.ac.uk> Subject: Re: Efficiency problem with mail queue => From: dwing@uh01.Colorado.EDU (Dan Wing) => How many outgoing SMTP agents have you configured? I'd recommend setting => up at *least* two for any MX configuration, and yours might benefit from => more -- if you have two, one of them can be busy sending a message to all => your 460 subscribers, and the other can be sending out other things at => the same time. => We have a large list which sometimes "jams" things up. So I set up another SMTP agent. This worked great until we had 2 messages to the same large list in the queue at the same time....and this is only a (very overworked) old MicroVAX II.. It would be nice if certain repeated agents could be reserved to short transactions only. Anything large, and it will have to wait until the 'main' smtp agent is free... Just a thought, Jamie. ================================================================================ Archive-Date: Mon, 15 Nov 1993 19:48:07 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Efficiency problem with mail queue Message-ID: <1993Nov16.022154.1@sejnet.sunet.se> From: eric@sejnet.sunet.se (Eric Thomas) Date: 16 Nov 93 02:21:54 WET Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17546956@MVB.SAIC.COM>, Jamie Jones <jamie@hicom.lut.ac.uk> writes: > We have a large list which sometimes "jams" things up. So I set up another > SMTP agent. This worked great until we had 2 messages to the same large > list in the queue at the same time....and this is only a (very overworked) > old MicroVAX II.. > > It would be nice if certain repeated agents could be reserved to short > transactions only. Anything large, and it will have to wait until the > 'main' smtp agent is free... Another thing is that there should be a configurable limit on the maximum amount of recipients per SMTP message. One reason is that most SMTP servers can only handle a limited amount of recipients (usually 500 or so), and another one is that you won't have a "mega message" taking up forever to process. Eric ================================================================================ Archive-Date: Tue, 16 Nov 1993 00:52:50 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 08:26:03 0200 From: Mircea Giurgiu <mgiurgiu@imar.ro> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <009759E2.B0437180.29849@imar.ro> Subject: Re: Efficiency problem with mail queue >I have a prelimiary copy of my "Tuning MX" document which I'll send you via >mail (it includes information on setting up multiple MX agents, and why >you'd want to setup multiple agents). > >-Dan Wing, Systems Administrator, University Hospital, Denver > dwing@uh01.colorado.edu or wing@eisner.decus.org > > Vote to sacrifice Ratbert! Send Email to scottadams@aol.com Send me one too!! Thanks, Mircea ------------------------------------------------------------------------------- Mircea Giurgiu, VAX system manager phones: +40.1.6.50.05.92 Institute of Mathematics of the +40.1.6.50.05.95 Romanian Academy fax : +40.1.3.12.98.29 PO.BOX 1-764, RO-70700 e-mail: mgiurgiu@imar.ro Cal. Grivitei, 21 mgiurgiu@roearn.ici.ac.ro ROMANIA ------------------------------------------------------------------------------ ================================================================================ Archive-Date: Tue, 16 Nov 1993 06:44:18 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 06:44:14 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <009759D4.76D7E4FC.3@ALPHA.WKU.EDU> Subject: Finally, Wingra is about to release Jnet for AXP Eric Thomas posted the following to some newsgroups. I'm posting it here, since it could affect some MX sites. MX V3.3 does not currently include support for Jnet on AXP simply because Jnet never supported AXP. I'm interested in finding one or more MX Jnet test sites if anyone here intends to run Jnet on an AXP. I anticipate no problems adding the Jnet support---the code compiles already, I just had no way to test it. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ------------------------------------------------------------------------------- Newsgroups: bit.listserv.jnet-l Subject: Planned release dates for JNET/AXP From: Eric Thomas <ERIC@SEARN.BITNET> Date: Thu, 11 Nov 1993 21:51:55 +0100 Good news! I got the following answer from Wingra, with permission to repost: ------------------------------------------------------------------------- X-Trace: by wingra.bitnet; Relayed; Thu, 11 Nov 1993 13:31:26 -0600 Content-Identifier: JNET under AXP -Reply Message-ID: <"NW-SMF:233EE22C01221100*"@MHS> X-Env-NW-SMF-Extension: MHS-id: 715AE22C81342ED9 MIME-Version: 1.0 Content-Type: Text/Plain; charset=DEC-Multinational Content-Transfer-Encoding: 7BIT Hop-Count: 2 Date: Thu, 11 Nov 1993 13:43 From: KOCHER@JMAIL1.mhs.wingra.bitnet Subject: JNET under AXP -Reply To: ERIC THOMAS <ERIC@SEARN.BITNET> cc: Faye Hammen <hammen@wingra.bitnet>, Scott Dicks <"Scott Dicks"@CCPOST.ccmail.wingra.bitnet> Dear Eric, I am in the process of drafting a press release announcing our Jnet/AXP release. Wingra's silence on this topic, other than our commitment to ship Alpha support by year end, was based our on caution to pre-announce without a realistic release date. Until we completed internal testing to verify Jnet/AXP meets our high quality standards I was not prepared to announce. Jnet/AXP will be ready for general shipment in December. Jnet TCP NJE/AXP will follow within 60 days. SNA NJE/AXP will follow the TCP NJE/AXP shipment. Wingra is recruiting field test sites for the link drivers. Eric, you are welcome to share this information with the Jnet-L list. Wingra thanks you for your patience and support. Wingra is evolving to a much more customer focused operation. We need and want your insights and advice as to want features and enhancements have the most impact and value to your operations. We look forward to working much more closely with the Jnet installed base. Regards, Barbara Kocher ================================================================================ Archive-Date: Tue, 16 Nov 1993 09:27:40 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 09:23:00 CST From: George Franklin <spectator@OEHL.brooks.af.mil> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU Message-ID: <009759EA.A4A364E0.26526@oehl.brooks.af.mil> Subject: Need Information I created a mail list in MCP using the DEFINE LIST command. It created the list but, I am having problems populating the list. I read in the Mailing List/File Server manual about the ADD and REMOVE commands. How do I access these commands. George Franklin Brooks AFB, TX ================================================================================ Archive-Date: Tue, 16 Nov 1993 09:49:39 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 09:48:01 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <009759EE.239F3217.6@ALPHA.WKU.EDU> Subject: RE: Need Information George Franklin <spectator@OEHL.brooks.af.mil> writes: > >I created a mail list in MCP using the DEFINE LIST command. It created >the list but, I am having problems populating the list. > >I read in the Mailing List/File Server manual about the ADD and REMOVE >commands. How do I access these commands. > You send the commands in the body of a mail message to the -Request address. For example, to add people to MX-List, I send the following command in the body of a mail message to mx-list-request@wkuvx1 : ADD "so-and-so" <user@node.edu> Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Tue, 16 Nov 1993 11:17:53 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 09:11:55 PDT From: Jim Wheeler # 916-978-5089 <jimww@sacto.mp.usbr.gov> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: jimww@sacto.mp.usbr.gov Message-ID: <009759E9.190EF1F2.4178@sacto.mp.usbr.gov> Subject: RE: Need Information >From: MX%"MX-List@WKUVX1.WKU.EDU" 16-NOV-1993 08:38:00.05 > >I created a mail list in MCP using the DEFINE LIST command. It created >the list but, I am having problems populating the list. > >I read in the Mailing List/File Server manual about the ADD and REMOVE >commands. How do I access these commands. > >George Franklin >Brooks AFB, TX George, I have just been going through this, here are some of the mistakes I made. when you define or mod a list be sure to save and reset/cluster The save MUST preceed the reset/cluster. (the reset tells the mx routines to reread the file ...) use mail to the <list>-request for the add remove etc commands. you can have multiple lines but be sure to put switches before the address or else the swithces become part of the address. I love the mcp review command good luck jimww ================================================================================ # James W. Wheeler # # # Regional Systems & WAN Manager # voice: 916 978-5089 fax: 916-978-5284 # # Mid Pacific Region # # # U.S. Bureau of Reclamation # inet: jimww@sacto.mp.usbr.gov # # 2800 Cottage Way MP1130 # # # Sacramento, CA 95825 # # ================================================================================ ================================================================================ Archive-Date: Tue, 16 Nov 1993 12:39:51 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 10:03:00 PST From: Mike Johnson (415) 594-3530 <mcjohnson@wizard.farinon.harris.com> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-LIST@WKUVX1.WKU.EDU CC: mcjohnson@wizard.farinon.harris.com Message-ID: <009759F0.3B8B0A20.18719@wizard.farinon.harris.com> Subject: Install of MX 3.2 with Multinet 3.2B installed on non-system disk We are moving from a VAX 6310 to a 4500 this weekend. As part of the move I am re-installing the software since I can not take any of the 6310's disk drives to the 4500. The 6310 had Multinet and MX installed on the system disk. On the 4500 I am installing the products on a non-system disk. Multinet (3.2B) was installed, but can not be started until I shutdown the 6310 Multinet processes (duplicate IP address). When I tried to install MX (3.2), an error occured when it tried to link against the "NETLIB_MULTINET_SHR" image. I am assuming the error occured because Multinet was not start, and not becuase it is installed on a non-system disk. My questions are: 1) Can MX work with Multinet, when Multinet is installed on a non-system disk. 2) Is anyone doing this. If so, did (do) you have any problems or special setups. Thanks for any assistance. Mike ------------------------------------------------------------------------------- Michael Johnson Harris Corp./Farinon Div. 1691 Bayport Ave., San Carlos, CA 94070-5307 (Voice) 415 594-3530 (FAX) 415 594-3777 (Internet) mcjohnson@wizard.farinon.harris.com ------------------------------------------------------------------------------ ================================================================================ Archive-Date: Tue, 16 Nov 1993 15:26:36 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 22:24:59 +0100 From: "Rok Vidmar, RCU Lj." <rok.vidmar@uni-lj.si> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975A57.E2BC563E.27206@uni-lj.si> Subject: RE: Need Information > >I read in the Mailing List/File Server manual about the ADD and REMOVE > >commands. How do I access these commands. > > > You send the commands in the body of a mail message to the -Request > address. For example, to add people to MX-List, I send the following > command in the body of a mail message to mx-list-request@wkuvx1 : > > ADD "so-and-so" <user@node.edu> Any chances to incorporate these commands into MCP? MX would thus become even more manager friendly. Regards, Rok Vidmar inet: rok.vidmar@uni-lj.si UCC, University of Ljubljana x.400: S=vidmar;G=rok;O=uni-lj;P=ac;A=mail;C=si Kardeljeva pl. 17 phone: +386 61 1686439 <-- Changed! 61000 Ljubljana fax: +386 61 1686358 <-- Changed! Slovenia ================================================================================ Archive-Date: Tue, 16 Nov 1993 15:48:01 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 16 Nov 1993 15:44:12 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975A1F.E5AD1C9E.12@ALPHA.WKU.EDU> Subject: RE: Need Information "Rok Vidmar, RCU Lj." <rok.vidmar@uni-lj.si> writes: > >> ADD "so-and-so" <user@node.edu> > > Any chances to incorporate these commands into MCP? MX would thus become >even more manager friendly. > Already planned, but no timeframe yet. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Tue, 16 Nov 1993 16:07:20 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Pointer to MX? Message-ID: <1993Nov16.171514.15114@decuk.uvo.dec.com> From: map@decuk.uvo.dec.com (Mike Procter) Date: Tue, 16 Nov 1993 17:15:14 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Hello, Can anyone tell me how I go about getting a copy of the MX software? Cheers, Mike ================================================================================ Archive-Date: Tue, 16 Nov 1993 16:07:50 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: MX_SMTP_SERVER_THREADS Message-ID: <1993Nov16.140648.373@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 16 Nov 93 14:06:48 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Anyone have any experience with changing the number of allowed *incoming* SMTP connections? Any recommendations for adjustments to the process limits/ quotas? I've set my system to allow 16 incoming connections, but it is difficult to test the "highest load" configuration when using simple scripts and forcing bounces to far-flung sites. I've asked Matt for his insights into adjusting this value. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ----- From the MX "Management Guide", MX V3.3: 5.2 SMTP_SERVER The SMTP server is a detached, multi-threaded process. You can specify how many threads the server should handle simultaneously by defining a logical name: $ DEFINE/SYSTEM/EXEC MX_SMTP_SERVER_THREADS n The value of n should range from 1 to 16. The default is 4. The SMTP server may require larger process quotas/limits if more than four threads are allowed. ================================================================================ Archive-Date: Tue, 16 Nov 1993 19:41:14 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Tuning MX V3.3 queue file Message-ID: <00975A27.6E24AD92@sacto.mp.usbr.gov> From: henrym@sacto.mp.usbr.gov Date: Wed, 17 Nov 1993 00:38:08 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <1993Nov10.121349.18075@vax.oxford.ac.uk>, daveh@vax.oxford.ac.uk (Dave Hastings, OUCS) writes: >> Tuning your MX queue file >> MX V3.3 >> >> Dan Wing >> dwing@uh01.colorado.edu >> >> >> Long, Boring Method >> ------------------- >> >> 4j. You'll see a comparison of using different bucket sizes, and the >> prompt "Key 0 Bucket Size": Press RETURN to accept the default >> value. >> <RETURN> > >MX seems to work better if the queue file's cluster size is divisible by the >bucket size (On our system the cluster size is 51 blocks and the bucket size is >17 blocks). > >> >> 4k. The FDL editor will then spit out an informational message >> about the depth of key 0, and other information. Press RETURN >> to continue. >> <RETURN> > >The index should be as flat as possible. > >Dave >-- > >David Hastings | "There's nothing wrong with my sense >VAX Systems Programmer | of reality - I have it thoroughly >Oxford University Computing Services| serviced once a fortnight" - DNA >daveh@vax.oxford.ac.uk | > 51 is a pretty big cluster size unless most of the files on that disk are really big. But I agree - I used the same technique on ISAM databases on TOPS10 on RP07's a few years back, making the cluster size the same as the number of sectors per track, and then making the ISAM index blocks multiples of those, and performance really, really took off. -HWM ---------- Henry W. Miller Assistant Systems and Network Manager U.S. Bureau of Reclamation, Mid Pacific Region 2800 Cottage Way MP1130 Sacramento, CA 95825 (916) 978-5108 ================================================================================ Archive-Date: Tue, 16 Nov 1993 19:41:33 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Install of MX 3.2 with Multinet 3.2B installed on non-system disk Message-ID: <00975A26.BEB26EB2@sacto.mp.usbr.gov> From: henrym@sacto.mp.usbr.gov Date: Wed, 17 Nov 1993 00:33:13 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17575904@MVB.SAIC.COM>, Mike Johnson (415) 594-3530 <mcjohnson@wizard.farinon.harris.com> writes: >We are moving from a VAX 6310 to a 4500 this weekend. As part of the move I am >re-installing the software since I can not take any of the 6310's disk drives to > the >4500. The 6310 had Multinet and MX installed on the system disk. On the 4500 I > am >installing the products on a non-system disk. Multinet (3.2B) was installed, > but can >not be started until I shutdown the 6310 Multinet processes (duplicate IP > address). >When I tried to install MX (3.2), an error occured when it tried to link against > the >"NETLIB_MULTINET_SHR" image. I am assuming the error occured because Multinet > was not >start, and not becuase it is installed on a non-system disk. > > >My questions are: > >1) Can MX work with Multinet, when Multinet is installed on a non-system disk. > >2) Is anyone doing this. If so, did (do) you have any problems or special > setups. > > >Thanks for any assistance. > >Mike > >------------------------------------------------------------------------------- >Michael Johnson >Harris Corp./Farinon Div. >1691 Bayport Ave., San Carlos, CA 94070-5307 > >(Voice) 415 594-3530 (FAX) 415 594-3777 >(Internet) mcjohnson@wizard.farinon.harris.com >------------------------------------------------------------------------------ > Mike, If I am reading your message correctly, the answers are: yes, yes and maybe. We are running both MultiNet and MX on a non-system disk, as government policy dictates that we do not install non-system vendor software on the system disk. MultiNet (or whatever TCP/IP transport you are running) needs to be installed and started, i.e., the system logicals defined, before MX can be installed correctly. Why don't you give the new node it's own IP address and name during the cutover period. I've used this scheme a number of times during conversion efforts. (however, an instance comes to mind, about a decade or so ago, when the new machine never came online for production after a number of months, because it would never run more than a few hours without faulting. But this was a non-DEC machine...) -HWM ---------- Henry W. Miller Assistant Systems and Network Manager U.S. Bureau of Reclamation, Mid Pacific Region 2800 Cottage Way MP1130 Sacramento, CA 95825 (916) 978-5108 ================================================================================ Archive-Date: Wed, 17 Nov 1993 02:13:20 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 17 Nov 1993 09:56:00 0200 From: Mircea Giurgiu <mgiurgiu@imar.ro> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975AB8.6B26DB60.30337@imar.ro> Subject: RE: Pointer to MX? >Hello, > >Can anyone tell me how I go about getting a copy of the MX software? >Cheers, >Mike Mike, You can find MX via anonymous ftp from ftp.spc.edu. You'll need: [.MACRO32.SAVESETS]mx033.zip You can get it via e-mail by sending the folowing command in the body of a mail message to FILESERV@WKUVX1.BITNET SEND mx033 ------------------------------------------------------------------------------- Mircea Giurgiu, VAX system manager phones: +40.1.6.50.05.92 Institute of Mathematics of the +40.1.6.50.05.95 Romanian Academy fax : +40.1.3.12.98.29 PO.BOX 1-764, RO-70700 e-mail: mgiurgiu@imar.ro Cal. Grivitei, 21 mgiurgiu@roearn.ici.ac.ro ROMANIA ------------------------------------------------------------------------------ ================================================================================ Archive-Date: Wed, 17 Nov 1993 08:43:27 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 17 Nov 1993 06:43:07 PDT From: Jim Wheeler # 916-978-5089 <jimww@sacto.mp.usbr.gov> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: jimww@sacto.mp.usbr.gov Message-ID: <00975A9D.796C15A2.4767@sacto.mp.usbr.gov> Subject: RE: Need Information > From: "Rok Vidmar, RCU Lj." <rok.vidmar@uni-lj.si> > Subject: RE: Need Information > > > >I read in the Mailing List/File Server manual about the ADD and REMOVE > > >commands. How do I access these commands. > > > > > You send the commands in the body of a mail message to the -Request > > address. For example, to add people to MX-List, I send the following > > command in the body of a mail message to mx-list-request@wkuvx1 : > > > > ADD "so-and-so" <user@node.edu> > > Any chances to incorporate these commands into MCP? MX would thus become > even more manager friendly. I second that - STRONGLY jimww > > Regards, > > Rok Vidmar inet: rok.vidmar@uni-lj.si > UCC, University of Ljubljana x.400: S=vidmar;G=rok;O=uni-lj;P=ac;A=mail;C=si > Kardeljeva pl. 17 phone: +386 61 1686439 <-- Changed! > 61000 Ljubljana fax: +386 61 1686358 <-- Changed! > Slovenia ================================================================================ # James W. Wheeler # # # Regional Systems & WAN Manager # voice: 916 978-5089 fax: 916-978-5284 # # Mid Pacific Region # # # U.S. Bureau of Reclamation # inet: jimww@sacto.mp.usbr.gov # # 2800 Cottage Way MP1130 # # # Sacramento, CA 95825 # # ================================================================================ ================================================================================ Archive-Date: Wed, 17 Nov 1993 08:47:14 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 17 Nov 1993 09:18:38 EST From: "Brian Tillman" <tillman@swdev.si.com> Reply-To: MX-List@WKUVX1.WKU.EDU To: WKUVX1.WKU.EDU!MX-list@esseye.si.com Message-ID: <00975AB3.33433F40.18129@swdev.si.com> Subject: RE: Pointer to MX? Mircea Giurgiu (mgiurgiu@imar.ro) writes: >You can find MX via anonymous ftp from ftp.spc.edu. >You'll need: > [.MACRO32.SAVESETS]mx033.zip Hm.. I found MX under [.MX.MX033]. Is it located in two directories? -- Brian tillman_brian@si.com ================================================================================ Archive-Date: Wed, 17 Nov 1993 09:08:08 CDT Sender: list-mgr@WKUVX1.WKU.EDU From: tminton@desire.wright.edu (Tim Minton - GCPL - Systems Analyst) Reply-To: MX-List@WKUVX1.WKU.EDU Subject: Rewrite and Path question Message-ID: <1993Nov17.082749.15532@desire.wright.edu> Date: 17 Nov 93 08:27:49 EST To: MX-List@WKUVX1.WKU.EDU Ok all you MX gurus, I am new to MX. I just started it last night and then shut it down because I was unable to receive mail to an address we were able to receive using Multinet's SMPT server. I know it should be fairly simple with rewrites and Path settings. (I just am having a hard time figuring out what order things should be in. Here are the particulars NODENAME GCDL01 FQDN gcdl01.gcpl.lib.oh.us I am able to receive mail with MX from the internet from to the FQDN, my only problem is our business cards have USERNAME@gcpl.lib.oh.us on them. Node GCDL01 is the only system on our net right now, so MX receives the message but does not deliver it because it doesn't know where to route the message. Any insight would be appreciated. Thanks Tim -- Tim Minton - Systems Analyst Greene County Public Library TMINTON@gcpl.lib.oh.us TMINTON@desire.wright.edu Xenia, Oh ================================================================================ Archive-Date: Wed, 17 Nov 1993 10:06:08 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 17 Nov 1993 11:05:47 EST From: Mighty Firebreather <dragon@nscvax.princeton.edu> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: twinton@desire.wright.edu Message-ID: <00975AC2.2B2661C0.18451@nscvax.princeton.edu> Subject: RE: Rewrite and Path question tminton@desire.wright.edu (Tim Minton - GCPL - Systems Analyst) writes: >Ok all you MX gurus, I am new to MX. I just started it last night and then >shut it down because I was unable to receive mail to an address we were able to >receive using Multinet's SMPT server. I know it should be fairly simple with >rewrites and Path settings. (I just am having a hard time figuring out what >order things should be in. > >Here are the particulars > >NODENAME GCDL01 >FQDN gcdl01.gcpl.lib.oh.us > >I am able to receive mail with MX from the internet from to the FQDN, my only >problem is our business cards have USERNAME@gcpl.lib.oh.us on them. Node >GCDL01 is the only system on our net right now, so MX receives the message but >does not deliver it because it doesn't know where to route the message. >Any insight would be appreciated. > Do you have: DEFINE PATH "gcpl.lib.oh.us" LOCAL In your CONFIG.MCP? ************************************************************************* * Here, there be dragons! * * dragon@nscvax.princeton.edu * * * * I'm job hunting. Any offers or leads will be appreciated. * * Thanks! * * Richard B. Gilbert * ************************************************************************* ================================================================================ Archive-Date: Wed, 17 Nov 1993 13:32:06 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 17 Nov 1993 13:30:47 CST From: "David Harrold, Milwaukee School Of Engineering" <harrold@barcly.acsd.msoe.edu> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <00975AD6.6CCEC360.13810@barcly.acsd.msoe.edu> Subject: SMTP MAIL From: Processing Hi everybody, I had a user tell me that he couldn't receive mail on our VMS machine, but the same message could get through to the Novell server. The problem seems to be the reverse-path supplied in the MAIL From: command sent from the remote system. The remote system sends: MAIL From:<dale!c_ewc> The Novell server (Mercury v1.10, Pegasus v3.01) accepts this address, but MX's SMTP gives the following error message: 501 Invalid Address: <dale!c_ewc> When I try this to my OSF/1 machine, it accepts the MAIL from: command, but the From: address looks like: dale!c_ewc@odo.msoe.edu. Reading (well, skimming) RFC821 and RFC822 leads me to believe that all SMTP servers should reject an address in the form <dale!c_ewc>. Apparently, I am wrong, but the big question is, how can I make MX accept this address? Specifics: MicroVAX 3600/VMS V5.5 UCX V2.0 MX V3.3 Any help, hints, or information would be greatly appreciated. Thanks, Dave Harrold ............................................................................... David Harrold harrold@picard.msoe.edu Milwaukee School Of Engineering Phone: 414-277-7286 Academic Computing Services FAX: 414-277-7465 1025 N. Broadway Street Milwaukee, WI 53202-3109 ............................................................................... "Someday utilities will be user friendly, not Unix" - Ehud Gavron ================================================================================ Archive-Date: Wed, 17 Nov 1993 14:46:37 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: SMTP MAIL From: Processing Message-ID: <1993Nov17.132440.387@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 17 Nov 93 13:24:40 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU York 12180-3590 USA |In article <17616399@MVB.SAIC.COM>, "David Harrold, Milwaukee School Of Engineering"<harrold@barcly.acsd.msoe.edu> writes: >Reading (well, skimming) RFC821 and RFC822 leads me to believe that all >SMTP servers should reject an address in the form <dale!c_ewc>. Apparently, >I am wrong, but the big question is, how can I make MX accept this >address? The above address is invalid. From the deep dark depths of the MX archives: >Archive-Date: Fri, 20 Sep 1991 09:28:27 EDT >To: mx-list@vms.ecs.rpi.edu >Date: 20 Sep 91 13:29:40 GMT >From: madison@mdmvs.ecs.rpi.edu (Matt Madison) >Sender: mx-list-request@vms.ecs.rpi.edu >Message-ID: <f0qn!nl@rpi.edu> >References: <1991Sep20.072504.17278@newshost.anu.edu.au> >Reply-To: madison@vms.ecs.rpi.edu >Subject: Re: MX barfs on nasty address from Easynet > >In article <1991Sep20.072504.17278@newshost.anu.edu.au>, >david@phys.anu.eduau (David Baldwin) writes: >>My DEC sales rep is trying to send me e-mail through the >>DECPA MRGATE Easynet-internet gateway but the resulting address >"From:" line looks like: >"snoc02::mrgate::"a1::gonzalez tony""@snoc02.enet.dec.com >which causes MX to bounce the mail with a "Protocol Error". > >Right. The quotes on the inside should be "quoted" - as in" > > "snoc02::mrgate::\"a1::gonzalez tony\""@snoc02.enet.dec.com > >>If I "telnet phys 25" and enter >>MAIL FROM:<"anyaddress that starts with a quote"> >>It gives "501 Invalid address: <....>" > >There are valid addresses that contain quotation marks. You just have >to get them right, syntactically. > >>Is there any way to get MX top just accept the "From:" line >>without trying to interpret it through RFC822 or whatever? > >No, I'm afraid not. The gateway from which the Internet mail is >originating should format the MAIL FROM: address properly (as should >all gateways :-). >-- >Matthew Madison, Systems Programmer | Internet: madison@vms.ecs.rpi.edu >Engineering Computing Services | Bitnet: MADISON@RPIECSVX >Rensselaer Polytechnic Institute | >Troy, New (While checking this out and re-reading the RFCs (I was having a similar problem a while ago, it appears that MX doesn't allow an envelope From address (MAIL FROM:) that is null ("<>"): 220 buckie.hsc.colorado.edu MX V3.3 VAX SMTP server ready at ... HELO FAKE 250 Name lookup failed for FAKE MAIL FROM: <> 501 Invalid address: <> QUIT 221 alpha.wku.edu Service closing transmission channel which 'should' be allowed.) -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Wed, 17 Nov 1993 15:09:45 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Rewrite and Path question Message-ID: <2ce20t$f7h@gap.cco.caltech.edu> From: carl@SOL1.GPS.CALTECH.EDU (Carl J Lydick) Date: 17 Nov 1993 20:36:45 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <1993Nov17.082749.15532@desire.wright.edu>, tminton@desire.wright.edu (Tim Minton - GCPL - Systems Analyst) writes: =Ok all you MX gurus, I am new to MX. I just started it last night and then =shut it down because I was unable to receive mail to an address we were able to =receive using Multinet's SMPT server. I know it should be fairly simple with =rewrites and Path settings. (I just am having a hard time figuring out what =order things should be in. = =Here are the particulars = =NODENAME GCDL01 =FQDN gcdl01.gcpl.lib.oh.us = =I am able to receive mail with MX from the internet from to the FQDN, my only =problem is our business cards have USERNAME@gcpl.lib.oh.us on them. Node =GCDL01 is the only system on our net right now, so MX receives the message but =does not deliver it because it doesn't know where to route the message. =Any insight would be appreciated. Have you tried: DEFINE PATH GCPL.LIB.OH.US LOCAL (make sure it's before the * path). Save the configuration file, and use the MCP RESET command to get the router to reload the configuration. -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Wed, 17 Nov 1993 15:28:27 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 17 Nov 1993 15:27:57 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975AE6.CAF54A4E.12@ALPHA.WKU.EDU> Subject: Re: SMTP MAIL From: Processing dwing@uh01.Colorado.EDU (Dan Wing) writes: > >(While checking this out and re-reading the RFCs (I was having a similar >problem a while ago, it appears that MX doesn't allow an envelope From >address (MAIL FROM:) that is null ("<>"): > > 220 buckie.hsc.colorado.edu MX V3.3 VAX SMTP server ready at ... > HELO FAKE > 250 Name lookup failed for FAKE > MAIL FROM: <> > 501 Invalid address: <> > QUIT > 221 alpha.wku.edu Service closing transmission channel > >which 'should' be allowed.) > Actually, the RFC states that there's no space after the ":" in "MAIL FROM:". MX accepts "MAIL FROM:<>". I added support for the space in MAIL FROM: to work with an old version of Charon---I obviously didn't apply that change to <>. But MX itself is correctly following the RFC. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Wed, 17 Nov 1993 15:32:03 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: RE: Rewrite and Path question -- Thanks Message-ID: <1993Nov17.144948.15565@desire.wright.edu> From: tminton@desire.wright.edu (Tim Minton - GCPL - Systems Analyst) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 17 Nov 93 14:49:48 EST To: MX-List@WKUVX1.WKU.EDU In article <17608985@MVB.SAIC.COM>, Mighty Firebreather <dragon@nscvax.princeton.edu> writes: > tminton@desire.wright.edu (Tim Minton - GCPL - Systems Analyst) writes: > >>Here are the particulars >> >>NODENAME GCDL01 >>FQDN gcdl01.gcpl.lib.oh.us >> >>I am able to receive mail with MX from the internet from to the FQDN, my only >>problem is our business cards have USERNAME@gcpl.lib.oh.us on them. Node >>GCDL01 is the only system on our net right now, so MX receives the message but >>does not deliver it because it doesn't know where to route the message. >>Any insight would be appreciated. >> > > Do you have: > DEFINE PATH "gcpl.lib.oh.us" LOCAL > Thanks for the info. I worked on the configuration today. I had that in my path statements only it was after "*" SMTP. That is why I was having problems. I redid all the define paths and it worked today. I am really impressed with the software though. > In your CONFIG.MCP? > > > ************************************************************************* > * Here, there be dragons! * > * dragon@nscvax.princeton.edu * > * * > * I'm job hunting. Any offers or leads will be appreciated. * > * Thanks! * > * Richard B. Gilbert * > ************************************************************************* -- Tim Minton - Systems Analyst Greene County Public Library TMINTON@gcpl.lib.oh.us TMINTON@desire.wright.edu Xenia, Oh ================================================================================ Archive-Date: Wed, 17 Nov 1993 16:16:09 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: SMTP MAIL From: Processing Message-ID: <1993Nov17.145237.390@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 17 Nov 93 14:52:37 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17620526@MVB.SAIC.COM>, "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> writes: >Actually, the RFC states that there's no space after the ":" in "MAIL >FROM:". MX accepts "MAIL FROM:<>". I added support for the space in >MAIL FROM: to work with an old version of Charon---I obviously didn't >apply that change to <>. But MX itself is correctly following the >RFC. I'll learn to read yet; just give me time. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Thu, 18 Nov 1993 04:48:28 CDT Sender: list-mgr@WKUVX1.WKU.EDU From: smith@amigo.uicc.com Reply-To: MX-List@WKUVX1.WKU.EDU Subject: [Q] Invalid "from" header? Message-ID: <1993Nov18.035022.1@amigo.uicc.com> Date: Wed, 17 Nov 1993 19:50:22 GMT To: MX-List@WKUVX1.WKU.EDU Can anyone tell me why we are getting this message? i.e. the X-MX-Warning? It appears that the From: address should have the magic "<>" surrounding it. Chillon is our mail router and I'm trying to determine whether the problem is with the MX config or something on the mail router... Thanks for any help. rgds -------------------------------------------------------------------------------- Return-Path: <@AMIGO.UICC.COM,@chillon.uicc.com,@mv.mv.com:CMU-OpenVMS-IP-Request@DRYCAS.CLUB.CC.CMU.EDU> Date: Wed, 17 Nov 93 22:17:48 GMT X-MX-Warning: Warning -- Invalid "From" header. From: @chillon.uicc.com:NAT@HIACPA0.HIA.NRC.CA (NATALIE PROWSE) Subject: How do I find out what version we are running? To: cmu-openvms-ip@DRYCAS.CLUB.CC.CMU.EDU Message-ID: <1993Nov17.221748.13025@nrcnet0.nrc.ca> ================================================================================ Archive-Date: Thu, 18 Nov 1993 04:59:01 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Thu, 18 Nov 1993 04:58:30 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00975B58.0670D36A.25@ALPHA.WKU.EDU> Subject: RE: [Q] Invalid "from" header? smith@amigo.uicc.com writes: > >Can anyone tell me why we are getting this message? > >i.e. the X-MX-Warning? > >It appears that the From: address should have the magic "<>" surrounding it. >Chillon is our mail router and I'm trying to determine whether the problem >is with the MX config or something on the mail router... > I don't believe that From: lines are supposed to include source routes, as this one is: >From: @chillon.uicc.com:NAT@HIACPA0.HIA.NRC.CA (NATALIE PROWSE) chillon appears to be rewriting the original From: line, supplying itself in the source route address. Sounds to me like a chillon problem. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Thu, 18 Nov 1993 08:19:36 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Yep, I'm confused... Message-ID: <753556009snx@daytona.anorad.com> From: mikel@daytona.anorad.com (Mikel Beck) Reply-To: MX-List@WKUVX1.WKU.EDU Date: Wed, 17 Nov 93 17:06:49 GMT To: MX-List@WKUVX1.WKU.EDU Here's what I've got right now: DECUS UUCP 2.0 link, over modem to PSI, getting mail & news. MX 3.3 processing mail on local node, passing mail around to other DECnet nodes on the net. I have three VAXes on the net, one which is running the above software, one which is barely used for mail, and one server for a PATHWORKS network. All three machines are running CMU-Tek TCP/IP in addition to DECnet. There are also 12 DECstations on the net, some of which are running DECnet, all of which are running TCP/IP. For some reason the guy who configured these machines decided NOT to use our registered domain name ( anorad.com ) for these machines and made up his own naming scheme. What I'd like to do is get SMTP running via MX to deliver mail to other nodes on the net, since all available machines are running TCP/IP and can receive mail this way ( or so I think ). I tried to set this up, but was not able to have SMTP going for in-house delivery and UUCP for incoming and outgoing mail. Is there some trick to this? If anybody is doing this with a similiar configuration ( CMU-Tek, DECUS UUCP & MX ), please let me know what you did to get it going. I'm not worried about the DECstations just yet, once I get the rest of it going I'll let the administrator for those machines worry about it. Any information would be appreciated. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Mikel Beck E-Mail: mikel@daytona.anorad.com Programmer/Analyst Warnat & Associates, Inc. ================================================================================ Archive-Date: Thu, 18 Nov 1993 13:52:26 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Thu, 18 Nov 1993 14:42:15 EST From: "Jonathan E. Hardis" <hardis@garnet.nist.gov> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: hardis@garnet.nist.gov Message-ID: <00975BA9.9315F740.28694@garnet.nist.gov> Subject: Insufficient Memory for a SHOW QUEUE I don't know how to interpret the error: %MX-F-MEMALLOC, error allocating memory from zone RCPTZONE -LIB-F-INSVIRMEM, insufficient virtual memory which appears part way through an MCP QUEUE SHOW /ALL (& /FULL) listing. Clues on the root cause of (or solutions for) this would be appreciated. On another topic, I run MX over T*K/CM* TCP/IP. As reported before, NAMRES tends to die when overworked after each reboot, a problem which can be cured for the duration by manually forcing it to EXIT and then by restarting it. Recently, as NAMRES went down for the count, MX seemed to do funny things to two queue entries, which now appear to be stuck. That is, they are both "In progress," but not doing anything. One is a locally generated message: Entry: 28191, Origin: [VMSmail] XXXXXXX Status: IN-PROGRESS, size: 193 bytes Created: 15-NOV-1993 08:55:55.37, expires 15-DEC-1993 08:55:55.37 Last modified 15-NOV-1993 08:58:57.02 Recipient #1: <xxxx@xxxxx.umd.edu> The other is message which was expanded by MLF: Entry: 28188, Origin: [Local] <Postmaster@garnet.nist.gov> Status: IN-PROGRESS, size: 19360 bytes Created: 15-NOV-1993 08:08:01.15, expires 15-DEC-1993 08:08:01.15 Last modified 15-NOV-1993 08:34:42.48 SMTP entry #28190, status: IN-PROGRESS, size: 18876 bytes Created: 15-NOV-1993 08:09:25.37, expires 15-DEC-1993 08:08:01.15 Last modified 15-NOV-1993 08:34:41.60 Recipient #1: ... Recipient #2: ... [etc.] This problem was not cured by restarting NAMRES or by rebooting. Subsequant mail works allright. Can anyone suggest a magic incantation, short of PURGE, to restart the processing of these queue entries? Thanks, Jonathan ================================================================================ Archive-Date: Thu, 18 Nov 1993 14:06:22 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Thu, 18 Nov 1993 14:03:02 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: hardis@garnet.nist.gov Message-ID: <00975BA4.18201129.8@ALPHA.WKU.EDU> Subject: RE: Insufficient Memory for a SHOW QUEUE "Jonathan E. Hardis" <hardis@garnet.nist.gov> writes: > >I don't know how to interpret the error: > >%MX-F-MEMALLOC, error allocating memory from zone RCPTZONE >-LIB-F-INSVIRMEM, insufficient virtual memory > >which appears part way through an MCP QUEUE SHOW /ALL (& /FULL) listing. > It means that there was insufficient virtual memory. 8-) This bug has already been fixed for the next release of MX. >Recently, as NAMRES went down for the count, MX seemed to do funny things >to two queue entries, which now appear to be stuck. That is, they are both >"In progress," but not doing anything. One is a locally generated >message: [...] >Can anyone suggest a magic incantation, short of PURGE, to restart the >processing of these queue entries? > Using the MCP QUEUE READY xxxx command should get them going again. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Thu, 18 Nov 1993 17:02:58 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Insufficient Memory for a SHOW QUEUE Message-ID: <2cgrnn$1a2@gap.cco.caltech.edu> From: carl@SOL1.GPS.CALTECH.EDU (Carl J Lydick) Date: 18 Nov 1993 22:07:51 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17657837@MVB.SAIC.COM>, "Jonathan E. Hardis" <hardis@garnet.nist.gov> writes: =I don't know how to interpret the error: = =%MX-F-MEMALLOC, error allocating memory from zone RCPTZONE =-LIB-F-INSVIRMEM, insufficient virtual memory = =which appears part way through an MCP QUEUE SHOW /ALL (& /FULL) listing. = =Clues on the root cause of (or solutions for) this would be appreciated. Well, your queue file could be hosed. You could try shutting down MX and using CONVERT (NOT CONVERT/RECLAIM) to create a new queue file. Or you would try MCP QUEUE PURGE before the MCP QUEUE SHOW/ALL. Or you could increase the page file quota for the process executing the command (and/or increase the SYSGEN parameter VIRTUALPAGECNT). =Recently, as NAMRES went down for the count, MX seemed to do funny things =to two queue entries, which now appear to be stuck. That is, they are both ="In progress," but not doing anything. One is a locally generated =message: = =Entry: 28191, Origin: [VMSmail] XXXXXXX = Status: IN-PROGRESS, size: 193 bytes = Created: 15-NOV-1993 08:55:55.37, expires 15-DEC-1993 08:55:55.37 = Last modified 15-NOV-1993 08:58:57.02 = Recipient #1: <xxxx@xxxxx.umd.edu> = =The other is message which was expanded by MLF: = =Entry: 28188, Origin: [Local] <Postmaster@garnet.nist.gov> = Status: IN-PROGRESS, size: 19360 bytes = Created: 15-NOV-1993 08:08:01.15, expires 15-DEC-1993 08:08:01.15 = Last modified 15-NOV-1993 08:34:42.48 = SMTP entry #28190, status: IN-PROGRESS, size: 18876 bytes = Created: 15-NOV-1993 08:09:25.37, expires 15-DEC-1993 08:08:01.15 = Last modified 15-NOV-1993 08:34:41.60 = Recipient #1: ... = Recipient #2: ... = [etc.] = =This problem was not cured by restarting NAMRES or by rebooting. =Subsequant mail works allright. = =Can anyone suggest a magic incantation, short of PURGE, to restart the =processing of these queue entries? Try: MCP> QUEUE READY 28191 MCP> QUEUE READY 28188 -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Thu, 18 Nov 1993 20:53:04 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: RE: [Q] Invalid "from" header? Message-ID: <1993Nov19.032238.1@sejnet.sunet.se> From: eric@sejnet.sunet.se (Eric Thomas) Date: 19 Nov 93 03:22:38 WET Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17643757@MVB.SAIC.COM>, "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> writes: > I don't believe that From: lines are supposed to include source > routes, as this one is: > >>From: @chillon.uicc.com:NAT@HIACPA0.HIA.NRC.CA (NATALIE PROWSE) In its infinite wisdom, RFC822 specifies that this address is indeed invalid, although From: <@chillon.uicc.com:NAT@HIACPA0.HIA.NRC.CA> (NATALIE PROWSE) From: <@chillon.uicc.com:NAT@HIACPA0.HIA.NRC.CA> From: NATALIE PROWSE <@chillon.uicc.com:NAT@HIACPA0.HIA.NRC.CA> would all be valid. Surprised? Eric ================================================================================ Archive-Date: Sun, 21 Nov 1993 14:45:14 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Sun, 21 Nov 1993 13:42:27 -0700 From: "Ray Harwood -- Data Basix: (602)721-1988" <rharwood@Data.Basix.COM> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-LIST@WKUVX1.WKU.EDU CC: rharwood@Data.Basix.COM Message-ID: <00975DFC.B7FF1B40.14897@Data.Basix.COM> Subject: No "destination-specific" queue entries I just encountered a strange MX-occurance... things are "fixed" now, but I was wondering how to avoid this in the future... East.Pima.EDU runs VMS 5.5-2, UCX 2.0, and MX 3.3. I hadn't received my "daily user's report" which is EMailed to me here at Data.Basix, so I logged into East.Pima.EDU and did an MCP QUE SHOW, and there were almost 100 entries, *none* of them with the usual second-line "destination-specific" entries. MCP READY <entry-num> had no effect. There were main entries for both MAIL and SMTP sources. In my haste to get things moving again, I probably didn't check out everything I could, but an MCP STATUS command listed all the "right stuff", and a SHOW SYSTEM showed all processes in a "normal" state. An MCP RESET had no effect, but an MCP SHUTDOWN followed by @SYS$SYSTEM:MCP_STARTUP caused things to flow normally. (1) Next time this happens, what else should I look at? (2) Does this situation -- with the limited information I managed to scrape together -- look familiar or indicate anything specific to look at? Thanks! Ray ----- Ray Harwood | Data Basix | Adjunct Faculty, East Campus, Voice: (602)721-1988 | PO Box 18324 | Pima Community College FAX: (602)721-7240 | Tucson, AZ 85731 | Instructor in Ada and Pascal rharwood@Data.Basix.COM | Info@Data.Basix.COM | rharwood@east.pima.edu ================================================================================ Archive-Date: Sun, 21 Nov 1993 18:01:02 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: . Message-ID: <1993Nov21.124057.12547@hadassah.bitnet> From: najman@hadassah.bitnet (Hadassah University Hospital. Jerusalem, Israel) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 21 Nov 93 12:40:56 +0200 To: MX-List@WKUVX1.WKU.EDU -- ================================================================================ Archive-Date: Sun, 21 Nov 1993 22:07:07 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: No "destination-specific" queue entries Message-ID: <2cpb0p$17l@gap.cco.caltech.edu> From: carl@SOL1.GPS.CALTECH.EDU (Carl J Lydick) Date: 22 Nov 1993 03:17:45 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17743602@MVB.SAIC.COM>, "Ray Harwood -- Data Basix: (602)721-1988" <rharwood@Data.Basix.COM> writes: =I just encountered a strange MX-occurance... things are "fixed" now, but I was =wondering how to avoid this in the future... = =East.Pima.EDU runs VMS 5.5-2, UCX 2.0, and MX 3.3. I hadn't received my "daily =user's report" which is EMailed to me here at Data.Basix, so I logged into =East.Pima.EDU and did an MCP QUE SHOW, and there were almost 100 entries, =*none* of them with the usual second-line "destination-specific" entries. = =MCP READY <entry-num> had no effect. There were main entries for both MAIL and =SMTP sources. = =In my haste to get things moving again, I probably didn't check out everything =I could, but an MCP STATUS command listed all the "right stuff", and a SHOW =SYSTEM showed all processes in a "normal" state. An MCP RESET had no effect, =but an MCP SHUTDOWN followed by @SYS$SYSTEM:MCP_STARTUP caused things to flow =normally. = =(1) Next time this happens, what else should I look at? = =(2) Does this situation -- with the limited information I managed to scrape =together -- look familiar or indicate anything specific to look at? This is just a guess, but I suspect that your "MX Router" process died or hung. It's the process responsible for deciding how to deliver a message. Should you encounter the problem again, use SHOW SYSTEM to find out what is and isn't running on your system. -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Mon, 22 Nov 1993 00:07:22 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Problem with MX_NAME_CONVERSION Message-ID: <1993Nov22.184421.22@galaxy> From: fowlerk@southpower.co.nz (Ken Fowler - Systems Administrator) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 22 Nov 93 18:44:21 NZDT To: MX-List@WKUVX1.WKU.EDU Hi, I am trying to get MX_NAME_CONVERSION to work but I am having problems. The system startup looks like this... > $ define/system/exec/nolog mx_site_name_conversion - > MX_EXE:MX_NAME_CONVERSION.EXE > $ define/system/exec/nolog mx_nicknames - > MX_EXE:MX_NAME_CONVERSION.DAT > $ define/system/exec/nolog mdm_timezone "+1200" > $ @sys$startup:mx_startup and the image is installed as follows... > GEMINI » install list mx_site_name_conversion/fu > > DISK$SYSTEMV5:<MX.EXE>.EXE > MX_NAME_CONVERSION;1 > Open Hdr Lnkbl > Entry access count = 3 > Current / Maximum shared = 2 / 2 > and I have the following in the nickname file... > GEMINI » ty mx_nicknames > BUXTONR Robert.Buxton > FOWLERK Ken.Fowler My problem is that when I try to send mail to mx%"Ken.Fowler" it bounces with a local delivery error. The bounced message looks like this... > From: MX%"Postmaster@southpower.co.nz" > To: SYSTEM > CC: > Subj: LOCAL delivery error > > Return-Path: <> > Date: Mon, 22 Nov 1993 18:33:49 +1200 > From: Local delivery agent <Postmaster@southpower.co.nz> > To: <system@southpower.co.nz> > Subject: LOCAL delivery error > X-Report-Type: Nondelivery; boundary="> Error description:" > > Note: this message was generated automatically. > > An error was detected while processing the enclosed message. A list of > the affected recipients follows. This list is in a special format that > allows software like LISTSERV to automatically take action on incorrect > addresses; you can safely ignore the numeric codes. > > --> Error description: > Error-For: ken$fowler@southpower.co.nz > Error-Code: 3 > Error-Text: No such local user > > Error-End: 1 error detected and the router debug file looks like this... > %PROCESS, Processing entry number 2022 > %PROCESS, Status from READ_INFO was 00000001 > %PROCESS, Message originated in VMS Mail. > %PROCESS, will run domain expander on envelope addresses. > %PROCESS, Processing address: <ken.fowler@southpower.co.nz> > %PROCESS, ... address now reads: <ken.fowler@southpower.co.nz> > %PROCESS, will run domain expander on message headers. > %PROCESS, Finished VMSmail-origin preprocessing. > %PROCESS, Recipient #0: <ken.fowler@southpower.co.nz> > %REWRITE, No rewrite rules matched <ken.fowler@southpower.co.nz> > %FINDPATH, Site-spec expand on southpower.co.nz err=00000000 > %FINDPATH, domain name SOUTHPOWER.CO.NZ matched path pattern SOUTHPOWER.CO.NZ > %PROCESS, Rewrote <ken.fowler@southpower.co.nz> as > <ken.fowler@southpower.co.nz> > - next hop southpower.co.nz, path 1 > %FINDALIAS, no alias found for ken.fowler > %PROCESS, no alias found for ken.fowler > %PROCESS, this is just a local delivery > %PROCESS, Adding to LOCAL path: ken.fowler. > %PROCESS, Path LOCAL gets 1 rcpts, entry number 2023 Does anyone know what is wrong with this?? Outgoing mail has the return address set as ken.Fowler@southpower.co.nz but trying to reply to the message does not work. Please send any email to fowlerk@southpower.co.nz rather than the address in the sig as I have disabled nicknames until I get it sorted out. Thanks, Ken. -- Southpower, Phone: +64 3 363-9527 Private Bag, Fax: +64 3 363-9816 Christchurch, Internet: Ken.Fowler@southpower.co.nz New Zealand. PSI: 0530130010083::fowler ================================================================================ Archive-Date: Mon, 22 Nov 1993 00:27:09 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 22 Nov 1993 00:29:13 CST From: Kenny Kon <kon@bible.acu.edu> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <00975E57.11E9A6C0.6137@bible.acu.edu> Subject: File server structure Date sent: 22-NOV-1993 00:24:41 If I have read the MX manual on file servers correctly, there can be only one level of subdirectories below the root directory pointed to by the fileserver. The anonymous FTP site here has subdirectories with more than one level, for e.g. the root directory is [ANONYMOUS], there are many subdirectories here like [ANONYMOUS.JAM], [ANONYMOUS.MISSIONS], etc. and these subdirectories have subdirectories as well. Do I have to create a whole new directory for the fileserver or is there a way round the limitation? Thanks. --------------------------------------------------------------------------- Kenny Kon Internet: kon@bible.acu.edu ACU Station, Box 7725 kkon@acuvax.acu.edu Abilene, TX 79699 BITNET: kkon@acuvax.bitnet (915)672-4915 CompuServe: 72724.3547@compuserve.com (915)674-3708 (work) ================================================================================ Archive-Date: Mon, 22 Nov 1993 01:24:54 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 22 Nov 1993 00:22:15 -0700 From: "Ray Harwood -- Data Basix: (602)721-1988" <rharwood@Data.Basix.COM> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: rharwood@Data.Basix.COM Message-ID: <00975E56.18962300.14955@Data.Basix.COM> Subject: RE: File server structure On Date: Mon, 22 Nov 1993 00:29:13 CST, Kenny Kon <kon@bible.acu.edu> writes: > If I have read the MX manual on file servers correctly, there can be only > one level of subdirectories below the root directory pointed to by the > fileserver. The anonymous FTP site here has subdirectories with more than > one level, for e.g. the root directory is [ANONYMOUS], there are many > subdirectories here like [ANONYMOUS.JAM], [ANONYMOUS.MISSIONS], etc. and > these subdirectories have subdirectories as well. Do I have to create a > whole new directory for the fileserver or is there a way round the > limitation? Thanks. I use the following, which only works when the fileserver and anonymous FTP directories are on the same device. Note: some of it is a little unconventional, but it works for me! (1) I have directory DUA2:[ANONYMOUS.PUB] set up for anonymous FTP access. (2) I have directory DUA2:[MD] as top-level fileserver directory for my muscular-dystrophy discussion group. In it, I've set up several sub-directories, such as [.ARTICLES], [.STORIES], [.TECHNO], etc. (3) I also have subdirectory DUA2:[MD.ARCHIVE] set up as the archive location for the MD listserver, and used the MCP command MOD LIST MD-List/ARCHIVE=DUA2:[MD.ARCHIVE]ARCHIVE so that files get named [MD.ARCHIVE]ARCHIVE.9311, for example. [Note: many folks put all the archives for all their lists in one directory; I prefer to have the archives for each list in a list-specific location.] (4) Now, to tie the [MD] directory underneath the [ANONYMOUS.PUB] directory, use the SET FILE/ENTER command: SET FILE DUA2:[000000]MD.DIR /ENTER=DUA2:[ANONYMOUS.PUB]MD.DIR Now, the anonymous FTP user sees the directory tree: [ANONYMOUS.PUB] [ANONYMOUS.PUB.MD] [ANONYMOUS.PUB.MD.ARTICLES] [ANONYMOUS.PUB.MD.ARCHIVES] while the MD-Files index shows all the "right stuff" from the DUA2:[MD.] root. The only (minor!) drawback is that the FTP users get to see all the ugly-looking file names which make the fileserver run so well. Ray ----- Ray Harwood | Data Basix | Adjunct Faculty, East Campus, Voice: (602)721-1988 | PO Box 18324 | Pima Community College FAX: (602)721-7240 | Tucson, AZ 85731 | Instructor in Ada and Pascal rharwood@Data.Basix.COM | Info@Data.Basix.COM | rharwood@east.pima.edu ================================================================================ Archive-Date: Mon, 22 Nov 1993 09:27:20 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: 22 Nov 93 15:22:30+0000 From: Fabien Marathee <marathee@eliot.cnes.fr> Reply-To: MX-List@WKUVX1.WKU.EDU Message-ID: <9311221522.AA14558@eliot.cnes.fr> To: MX-List@WKUVX1.WKU.EDU Subject: MX security CC: spiteri@melies.cnes.fr It seems there are several security holes in sendmail programs on Unix systems, depending on the version. Are there any 'known' problems of the same kind on VMS systems running MX ? Are there any MX configuration recommandations concerning security (SMTP/IP network connection) ? Can I retrict SMTP access to know SMTP servers with MX (I am running Multinet) ? Thank you. Fabien Marathee, marathee@eliot.cnes.fr ================================================================================ Archive-Date: Mon, 22 Nov 1993 10:02:26 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 22 Nov 1993 09:03:06 MST From: "Louis B. Moore" <MOORE.LOUIS@tchden.org> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU, fowlerk@southpower.co.nz Message-ID: <00975E9E.DBA6E260.16687@tchden.org> Subject: RE: Problem with MX_NAME_CONVERSION > GEMINI » ty mx_nicknames > BUXTONR Robert.Buxton > FOWLERK Ken.Fowler Try setting up your nicknames to be: FOWLERK KEN.FOWLER I convert the address to uppercase to do the look up. Louis B. Moore Internet: MOORE.LOUIS@tchden.org Systems Programmer Phone: +1.303.837.2513 The Children's Hospital of Denver Denver, Colorado USA 80218 ================================================================================ Archive-Date: Mon, 22 Nov 1993 11:47:33 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: File server structure Message-ID: <1993Nov22.084919.419@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 22 Nov 93 08:49:19 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17754160@MVB.SAIC.COM>, Kenny Kon <kon@bible.acu.edu> writes: >If I have read the MX manual on file servers correctly, there can be only >one level of subdirectories below the root directory pointed to by the >fileserver. The anonymous FTP site here has subdirectories with more than >one level, for e.g. the root directory is [ANONYMOUS], there are many >subdirectories here like [ANONYMOUS.JAM], [ANONYMOUS.MISSIONS], etc. and >these subdirectories have subdirectories as well. Do I have to create a >whole new directory for the fileserver or is there a way round the >limitation? Thanks. You could put the anonymous files and the MX fileserver on the same disk and use SET FILE/ENTER to avoid having duplicate copies of the files taking space on your disk. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org Vote to sacrifice Ratbert! Send Email to scottadams@aol.com ================================================================================ Archive-Date: Mon, 22 Nov 1993 13:00:09 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: MX security Message-ID: <1993Nov22.111013.423@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 22 Nov 93 11:10:12 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17767782@MVB.SAIC.COM>, Fabien Marathee <marathee@eliot.cnes.fr> writes: >It seems there are several security holes in sendmail programs on Unix >systems, depending on the version. Are there any 'known' problems of the same >kind on VMS systems running MX ? Almost all of the holes in sendmail are a result of being able to get the daemon (running on the server machine) to execute a script. None of the VMS-based implementations of SMTP servers (MX, PMDF, MultiNet, TCPware, Wollongong, etc.) are implemented like Unix's sendmail -- none of them have any of the holes present in sendmail. *If* any of the VMS-based SMTP implementations have holes, the means to break into them would be completely different than the means used to break into sendmail SMTP servers. >Are there any MX configuration recommandations concerning security >(SMTP/IP network connection) ? > >Can I retrict SMTP access to know SMTP servers with MX (I am running >Multinet) ? Why would you want to? If you did this, you couldn't, for example, receive this message I'm CC'ing to you, because you wouldn't have setup my system to send mail to your system. And, no, you cannot do this with MX. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Mon, 22 Nov 1993 13:14:14 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: [Q] VMS mail patch? Message-ID: <1993Nov22.121921.1@amigo.uicc.com> From: smith@amigo.uicc.com Reply-To: MX-List@WKUVX1.WKU.EDU Date: Mon, 22 Nov 1993 04:19:21 GMT To: MX-List@WKUVX1.WKU.EDU Some time back I saw a reference to a patch for VMS mail that allowed the entry of internet addresses without the MX prefix... Can anyone point me in the right direction for this? I am looking for VAX/VMS 5.5-2 and shortly AXP VMS 1.5 if it makes a difference (which I am sure it does). Is there an FAQ for this list? I used to be on the MX mailing list be haven't had any mail for some time. Does it still exist? rgds ================================================================================ Archive-Date: Mon, 22 Nov 1993 13:43:06 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: [Q] VMS mail patch? Message-ID: <1993Nov22.123101.424@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 22 Nov 93 12:31:00 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <1993Nov22.121921.1@amigo.uicc.com>, smith@amigo.uicc.com writes: >Some time back I saw a reference to a patch for VMS mail that allowed >the entry of internet addresses without the MX prefix... Can anyone >point me in the right direction for this? FTP to ftp.spc.edu, and look in the [.MX] directory. An older version (which doesn't work with VMS V6.0) is in the [.CONTRIB] directory of the MX V3.3 kit, if you chose to install contributed software when you installed MX V3.3. >I am looking for VAX/VMS 5.5-2 and shortly AXP VMS 1.5 if it makes a >difference (which I am sure it does). I don't know if there is a patch for AXP/VMS -- this is because AXP/VMS doesn't include the PATCH command. However, Hunter tweaked things a lot to get similar functionality for his DCL_RECALL program (so it would work with AXP/VMS), but I haven't heard of anyone trying to do the same thing to the AXP/VMS MAILSHR.EXE image. I imagine you could, maybe, make the patch on a VAX and then VEST it over to an AXP....? >Is there an FAQ for this list? > >I used to be on the MX mailing list be haven't had any mail for some time. >Does it still exist? Matt Madison now works for TGV, and MX is now maintained by Hunter Goatley. The MX mailing list is now run from Hunter's site -- MX-LIST@WKU.EDU. Send a mail message to LISTSERV@WKU.EDU (or MXSERVER@WKU.EDU) to subscribe to MX-LIST. The mailing list and the newsgroup (vmsnet.mail.mx) are gatewayed to each other. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Mon, 22 Nov 1993 15:35:04 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: File server structure Message-ID: <2cr9av$se2@gap.cco.caltech.edu> From: carl@SOL1.GPS.CALTECH.EDU (Carl J Lydick) Date: 22 Nov 1993 21:01:19 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17754160@MVB.SAIC.COM>, Kenny Kon <kon@bible.acu.edu> writes: =Date sent: 22-NOV-1993 00:24:41 = =If I have read the MX manual on file servers correctly, there can be only =one level of subdirectories below the root directory pointed to by the =fileserver. The anonymous FTP site here has subdirectories with more than =one level, for e.g. the root directory is [ANONYMOUS], there are many =subdirectories here like [ANONYMOUS.JAM], [ANONYMOUS.MISSIONS], etc. and =these subdirectories have subdirectories as well. Do I have to create a =whole new directory for the fileserver or is there a way round the =limitation? Thanks. You could create a new directory for the fileserver, then enter each of the subdirectories in it: $ CREATE/DIR [FILESERVER] $ SET FILE/ENTER=[FILESERVER]JAM.DIR;1 [ANONYMOUS]JAM.DIR;1 $ SET FILE/ENTER=[FILESERVER]PEACH.DIR;1 [ANONYMOUS.JAM]PEACH.DIR;1 and so forth. -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Tue, 23 Nov 1993 09:21:48 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 23 Nov 1993 09:19:57 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <00975F6A.60599D9C.21@ALPHA.WKU.EDU> Subject: MX BOF at DECUS SF For those of you attending DECUS in San Francisco, there will be an MX BOF on Monday, December 6, 1993, at 4PM. I won't know which room until I get there, but it has been scheduled for the hour from 4PM--5PM. Matt Madison and I will both be there to discuss MX problems and future directions---yes, there will be a new version *sometime*! 8-) (In case you don't know, a BOF is a "Birds-of-Feather" session where you can sit around and talk about whatever common interest you have---MX, in this case.) Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Tue, 23 Nov 1993 11:46:31 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: 23 Nov 93 11:39:00 CST From: "John Clement-Rice U. (713)527-8101x2037" <CLEMENT@physics.rice.edu> Reply-To: MX-List@WKUVX1.WKU.EDU Subject: Re: MX security To: "MX-List" <MX-List@WKUVX1.WKU.EDU> The latest CMUIP (unreleased) currently has security to allow restrictions on usage of remote and local ports. Any remote port may be declared restricted, so that the process needs PHY_IO privilege. This allows you to prevent users from writing their own programs to access these ports on remote hosts. In addition local ports below 1024 can be resticted as a group, and any local port with a WKS definition can also be restricted. The main problem you might have is that a local user can access the SMTP remote port and send mail masquerading as a another local user. Incoming mail should not be capable of interception if you have a Daemon running, on the SMTP local port. The current version of CMU provides restrictions to prevent interception of incoming mail as well as preventing "Trojan horse mail" Now of course you can not be sure of the sender, when you get a message, unless you KNOW the remote host has good security. Another approach to security is usage of the IDENT service (also in latest CMU) to "know" who is doing what. This makes perpetrators more evident. The issues of security on the internet is a can of worms that is not really in my opinion solved, as of yet. Just be aware that when you get a missive from someone else, that it is like a telephone call. You need to identify the caller, and sometimes it is hard to do that, before you give out info such as credit card numbers etc. I personally would like to see an IDENT service built into TCP and UDP so that messages would include internal identification of the sender. This would allow private identification to the other end of the link without allowing anyone to find out is doing what. This would be similar to Tel. Co. Caller ID. The built in IDENT service could even be encrypted, or better yet have both plain text and encrypted parts. John Clement Rice University ================================================================================ Archive-Date: Tue, 23 Nov 1993 18:03:56 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Tue, 23 Nov 1993 18:03:05 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <00975FB3.7500C141.4@ALPHA.WKU.EDU> Subject: MX BOF at DECUS SF For those of you attending DECUS in San Francisco, a one-hour MX BOF (Birds-of-Feather) has been scheduled for Monday, December 6, 1993, from 4PM to 5PM. Matt Madison and I will both be there to discuss MX problems and future directions, or just to shoot the breeze with whoever else shows up. If you're gonna be in SF, then make plans to attend the BOF! Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Wed, 24 Nov 1993 16:30:40 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Ordering rewrite rules Message-ID: <1993Nov24.155957.12552@hadassah.bitnet> From: najman@hadassah.bitnet (Hadassah University Hospital. Jerusalem, Israel) Reply-To: MX-List@WKUVX1.WKU.EDU Date: 24 Nov 93 15:59:56 +0200 To: MX-List@WKUVX1.WKU.EDU Hi. How do I insert a rewrite rule (or path) in a particular order of an existing list instead of appending it at the tail. I have RTFMed, but can't find it. -- +----------+------------------+------------------------------+---------------+ ! !\ ! / ! NAJMAN KAHANA ! Hadassah University Hospital ! thanks, ! ! ! \ !/\ ! najman@HADASSAH ! Jerusalem, Israel ! we have our ! ! ! \! \ ! ! (visit our capital soon) ! own viruses ! +----------+------------------+------------------------------+---------------+ ================================================================================ Archive-Date: Wed, 24 Nov 1993 18:16:12 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 24 Nov 1993 18:38:49 EST From: Ian Vaz <ian@mars.senecac.on.ca> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-LIST@WKUVX1.WKU.EDU Message-ID: <00976081.9DFD9B40.9437@mars.senecac.on.ca> Subject: List server add file problem. Hi, I've setup a list server on a microvax. All is fine. However, when I subscribe to the list from a different VAX, I never get the ADD FILE, or the "you have been added to the list" message file mailed to me. I do get all postings to the list. When on the machine where the list resides, the add file is mailed to all users when they subscribe. Thus, to summarize. The add file is sent only if to users who subscribe to the list when they reside on the machine where the list has been created. However, no add file to users subscribing to the list from other machines. The users are added to the list and receive mailings in all cases. The list protections are: (S:RWED, O:RWED, G:RE, W:RE) All files and directories have WORLD READ and EXECUTE privs. Any ideas as to why the specified users are not getting the ADD FILE text sent to them? Thanks for any replies. Regards, Ian Vaz, Seneca College, Toronto, Canada, Ph:(416) 491-5050, Ext. 7252 e-mail: ian@mars.senecac.on.ca, ian@phobos.senecac.on.ca ================================================================================ Archive-Date: Wed, 24 Nov 1993 18:38:00 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Wed, 24 Nov 1993 17:21:22 -0700 From: "Ray Harwood -- Data Basix: (602)721-1988" <rharwood@Data.Basix.COM> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: rharwood@Data.Basix.COM Message-ID: <00976076.CBBEDE00.15446@Data.Basix.COM> Subject: RE: Ordering rewrite rules > How do I insert a rewrite rule (or path) in a particular order > of an existing list instead of appending it at the tail. > I have RTFMed, but can't find it. In my experience, "you can't" is the correct answer. Basically, I did an MCP SHOW PATH */COMMAND/OUTPUT=PATH.MCP, then edited the file to include REMOVE PATH XXXXX commands, and then DEFINE PATH XXXX commands, followed by a RESET... then go into MCP and do a @PATH.MCP to remove/add/reset the paths all in one swell foop. Anything easier, I'd like to know of too. It would be REALLY nice, but asking too much, I'm sure, to have a TPU interface such as MCP/EDIT/PATH comand, kindof like EDIT/ACL! ;)} Ray ----- Ray Harwood | Data Basix | Adjunct Faculty, East Campus, Voice: (602)721-1988 | PO Box 18324 | Pima Community College FAX: (602)721-7240 | Tucson, AZ 85731 | Instructor in Ada and Pascal rharwood@Data.Basix.COM | Info@Data.Basix.COM | rharwood@east.pima.edu ================================================================================ Archive-Date: Wed, 24 Nov 1993 23:25:45 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: Ordering rewrite rules Message-ID: <2d1d9j$69f@gap.cco.caltech.edu> From: carl@SOL1.GPS.CALTECH.EDU (Carl J Lydick) Date: 25 Nov 1993 04:45:39 GMT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <1993Nov24.155957.12552@hadassah.bitnet>, najman@hadassah.bitnet (Hadassah University Hospital. Jerusalem, Israel) writes: =Hi. = How do I insert a rewrite rule (or path) in a particular order =of an existing list instead of appending it at the tail. = = I have RTFMed, but can't find it. You delete all later rewrite rules, then add the rewrite rules in order. The easiest way to accomplish this is: $ MCP :== $MX_EXE:MCP $ MCP MCP> SHOW ALL/COMMAND/OUTPUT=MX_DIR:CONFIG.MCP MCP> EXIT {edit config.mcp to specify the desired rewrite rules, in the appropriate order} $ MCP/NOFILE MCP> @MX_DIR:CONFIG.MCP MCP> SAVE MX_DIR:MX_CONFIG.MXCFG MCP> RESET MCP> EXIT This also has the advantage that you've got the configuration stored in CONFIG.MCP. -------------------------------------------------------------------------------- Carl J Lydick | INTERnet: CARL@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My understanding of astronomy is purely at the amateur level (or below). So unless what I'm saying is directly related to VAX/VMS, don't hold me or my organization responsible for it. If it IS related to VAX/VMS, you can try to hold me responsible for it, but my organization had nothing to do with it. ================================================================================ Archive-Date: Mon, 29 Nov 1993 06:24:58 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 29 Nov 1993 13:15:48 CET From: Hans-Joachim Koch <sys_hjk@lifra.lif.de> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU CC: sys_hjk@lifra.lif.de Message-ID: <00976442.518B1E80.6659@lissy1.lif.de> Subject: Problem moving MX to other node Hello, we have a vaxcluster with 3 nodes. Only one of them has the multinet license. We had MX and Multinet on node LIFRA. Because we want to switch off this old 8530, we moved multinet to node LISSY1 and did so with MX. I checked all logicals, the whole MX configuration a couple of times... one problem is still there. The following mail was send from node LISSY1 (I'm logged on there now), but MX created still a header with the old node name LIFRA. > Return-Path: <sys_hjk@lifra.lif.de> ***** > Received: from lif.de (sunny.lif.de) by lissy1.lif.de (MX V3.3 VAX) with SMTP; > Mon, 29 Nov 1993 13:04:39 CET > Received: from lissy1.lif.de by lif.de (4.1/SMI-4.1) id AA06196; Mon, 29 Nov 93 > 13:09:14 +0100 > Received: by lissy1.lif.de (MX V3.3 VAX) id 6653; Mon, 29 Nov 1993 13:04:27 CET > Date: Mon, 29 Nov 1993 13:04:22 CET > From: Hans-Joachim Koch <sys_hjk@lifra.lif.de> ***** > To: koch@lif.de > CC: sys_hjk@lifra.lif.de ***** > Message-ID: <00976440.B938C8E0.6653@lissy1.lif.de> > Subject: test Any hints? Thanks for your help. Hans. -- Hans-Joachim Koch, Computer department of Lahmeyer International Lyoner Strasse 22, D-60528 Frankfurt, Germany Phone: +49 (69) 6677-642, Fax: +49 (69) 6677-388, Tx: 413478 li d ================================================================================ Archive-Date: Mon, 29 Nov 1993 06:37:46 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 29 Nov 1993 06:37:24 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU CC: sys_hjk@lissy1.lif.de Message-ID: <0097640A.A9D3B12F.7@ALPHA.WKU.EDU> Subject: RE: Problem moving MX to other node Hans-Joachim Koch <sys_hjk@lifra.lif.de> writes: > >Hello, > >we have a vaxcluster with 3 nodes. Only one of them has the multinet >license. We had MX and Multinet on node LIFRA. Because we want to switch >off this old 8530, we moved multinet to node LISSY1 and did so with MX. > >I checked all logicals, the whole MX configuration a couple of times... >one problem is still there. The following mail was send from node >LISSY1 (I'm logged on there now), but MX created still a header with the >old node name LIFRA. > I suspect that it's somewhere in the domain lookup. Without seeing an MX_ROUTER_LOG.LOG, I don't know for sure what it's doing (you can enable Router debugging with: $ DEFINE/SYS/EXEC MX_ROUTER_DEBUG TRUE >Any hints? Thanks for your help. The MX record defined for LISSY1 is not right: LISSY1.LIF.DE preference = 10, mail exchanger = sunny.lif.de.lif.de When trying an NSLOOKUP on it: *** wkuvx1.wku.edu can't find SUNNY.LIF.DE.LIF.DE: Non-existent domain I'll bet that's related to the problem. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Mon, 29 Nov 1993 11:26:19 CDT Sender: list-mgr@WKUVX1.WKU.EDU From: chen@wave.aoml.erl.gov (Ying-Yuang Chen) Reply-To: MX-List@WKUVX1.WKU.EDU Subject: Best full screen mail reader ? Date: 29 Nov 1993 11:33:04 -0500 Message-ID: <2dd880$lpb@wave.aoml.erl.gov> To: MX-List@WKUVX1.WKU.EDU Hi, Could someone tell me what is the best full screen mail reader for vms 5.5-2 ? and where to get it? We want to use it to replace all-in-one. -- Ying-Yuang Chen Computer Services, UNIX System Admin. US Department of Commerce NOAA / AOML 4301 Rickenbacker Causeway, Miami, FL 33199 ================================================================================ Archive-Date: Mon, 29 Nov 1993 11:47:53 CDT Sender: list-mgr@WKUVX1.WKU.EDU Subject: Re: List server add file problem. Message-ID: <1993Nov29.102045.437@buckie.hsc.colorado.edu> From: dwing@uh01.Colorado.EDU (Dan Wing) Date: 29 Nov 93 10:20:45 MDT Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU In article <17856385@MVB.SAIC.COM>, Ian Vaz <ian@mars.senecac.on.ca> writes: > (S:RWED, O:RWED, G:RE, W:RE) > >All files and directories have WORLD READ and EXECUTE privs. > >Any ideas as to why the specified users are not getting the >ADD FILE text sent to them? Thanks for any replies. You'd have to enable MLF debugging, and possibly Router debugging, to determine the problem. -Dan Wing, Systems Administrator, University Hospital, Denver dwing@uh01.colorado.edu or wing@eisner.decus.org ================================================================================ Archive-Date: Mon, 29 Nov 1993 11:58:17 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 29 Nov 1993 11:57:54 CST From: "Hunter Goatley, WKU" <goathunter@ALPHA.WKU.EDU> Reply-To: MX-List@WKUVX1.WKU.EDU To: MX-List@WKUVX1.WKU.EDU Message-ID: <00976437.6FF3A677.3@ALPHA.WKU.EDU> Subject: Re: List server add file problem. dwing@uh01.Colorado.EDU (Dan Wing) writes: > >In article <17856385@MVB.SAIC.COM>, Ian Vaz <ian@mars.senecac.on.ca> writes: >> (S:RWED, O:RWED, G:RE, W:RE) >> >>All files and directories have WORLD READ and EXECUTE privs. >> >>Any ideas as to why the specified users are not getting the >>ADD FILE text sent to them? Thanks for any replies. > >You'd have to enable MLF debugging, and possibly Router debugging, to >determine the problem. > This problem was resolved, but the solution wasn't posted here. Ian was running MX V3.1C, which did not form MAIL FROM: lines correctly if there were no system users defined in MCP. He upgraded to MX V3.3 and all is well again. Hunter ------ Hunter Goatley, VMS Systems Programmer, Western Kentucky University goathunter@ALPHA.WKU.EDU (or goathunter@WKUVX1.BITNET) ================================================================================ Archive-Date: Mon, 29 Nov 1993 13:09:40 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 29 Nov 1993 14:04:41 CST From: Ian Vaz <ian@mars.senecac.on.ca> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@wkuvx1.wku.edu Message-ID: <00976449.25FDF6A0.12065@mars.senecac.on.ca> Subject: List server add file problem-Thanks Hunter. Hi, Regarding the problem with MLF not sending the ADD file to newly subscribed users: I did have debugging enabled. However, the log file did not indicate any problems. Cut a long story short, I communicated with Hunter Goately on the issue and was accordingly chastised ;-) for running MX v3.1C and not the latest version, v3.3. After upgrading the issue has been resolved. Thanks Hunter. Thanks also for all the help from this list. Ian Vaz, Seneca College, Toronto, Canada, Ph:(416) 491-5050, Ext. 7252 e-mail: ian@mars.senecac.on.ca, ian@phobos.senecac.on.ca ================================================================================ Archive-Date: Tue, 30 Nov 1993 08:14:37 CDT Sender: list-mgr@WKUVX1.WKU.EDU Date: Mon, 29 Nov 1993 21:30:59 CET From: Hans-Joachim Koch <sys_hjk@lifra.lif.de> Reply-To: MX-List@WKUVX1.WKU.EDU To: mx-list@WKUVX1.WKU.EDU CC: sys_hjk@lifra.lif.de Message-ID: <00976487.7F2988A0.6711@lissy1.lif.de> Subject: Problem moving MX to other node - more details Hello, thanks Hunter, for the first hints. I hope, our nameserver guru will fix the problem tomorrow, I can't do it. To the naming problem. I very very strange. I send a mail to a mail reflector mirror@germany.eu.net. Debug was enabled for router. ---------------------------------------------------------------------------- 29-NOV-1993 21:11:32.15 %PROCESS, Processing entry number 6701 29-NOV-1993 21:11:32.70 %PROCESS, Status from READ_INFO was 00000001 29-NOV-1993 21:11:32.70 %PROCESS, Message originated in VMS Mail. 29-NOV-1993 21:11:32.71 %PROCESS, will run domain expander on envelope addresses. 29-NOV-1993 21:11:32.71 %PROCESS, Processing address: <mirror@germany.eu.net> 29-NOV-1993 21:11:35.65 %PROCESS, ... address now reads: <mirror@germany.eu.net> 29-NOV-1993 21:11:35.65 %PROCESS, will run domain expander on message headers. 29-NOV-1993 21:11:37.14 %PROCESS, Finished VMSmail-origin preprocessing. 29-NOV-1993 21:11:37.14 %PROCESS, Recipient #0: <mirror@germany.eu.net> 29-NOV-1993 21:11:37.14 %REWRITE, No rewrite rules matched <mirror@germany.eu.net> 29-NOV-1993 21:11:37.22 %FINDPATH, Site-spec expand on germany.eu.net err=00000870 29-NOV-1993 21:11:37.22 %FINDPATH, domain name GERMANY.EU.NET matched path pattern * 29-NOV-1993 21:11:37.22 %PROCESS, Rewrote <mirror@germany.eu.net> as <mirror@germany.eu.net> - next hop SUNNY.LIF.DE, path 2 29-NOV-1993 21:11:37.36 %PROCESS, Adding to SMTP path: <mirror@germany.eu.net>. 29-NOV-1993 21:11:37.97 %PROCESS, Path SMTP gets 1 rcpts, entry number 6702 ---------------------------------------------------------------------------- Thats not very much information (for me, at least), so I tried again and enabled smtp debug. ---------------------------------------------------------------------------- 29-NOV-1993 21:16:03.87 Processing queue entry number 6708 on node LISSY1 29-NOV-1993 21:16:04.45 Recipient: <mirror@germany.eu.net>, route=SUNNY.LIF.DE 29-NOV-1993 21:16:04.45 SMTP_SEND: looking up host name SUNNY.LIF.DE 29-NOV-1993 21:16:04.47 SMTP_SEND: Attempting to start session with SUNNY.LIF.DE [149.233.0.11] 29-NOV-1993 21:16:04.48 SMTP_SEND: Connected 29-NOV-1993 21:16:04.82 SMTP_SEND: Rcvd: 220 lif.de Sendmail 4.1/SMI-4.1 ready at Mon, 29 Nov 93 21:20:44 +0100 29-NOV-1993 21:16:04.96 SMTP_SEND: Sent: HELO lissy1.lif.de 29-NOV-1993 21:16:04.97 SMTP_SEND: Rcvd: 250 lif.de Hello lissy1.lif.de, pleased to meet you 29-NOV-1993 21:16:04.97 SMTP_SEND: Sent: MAIL FROM:<sys_hjk@lissy1.lif.de> 29-NOV-1993 21:16:05.00 SMTP_SEND: Rcvd: 250 <sys_hjk@lissy1.lif.de>... Sender ok 29-NOV-1993 21:16:05.00 SMTP_SEND: Sent: RCPT TO:<mirror@germany.eu.net> 29-NOV-1993 21:16:05.09 SMTP_SEND: Rcvd: 250 <mirror@germany.eu.net>... Recipient ok 29-NOV-1993 21:16:05.11 SMTP_SEND: Sent: DATA 29-NOV-1993 21:16:05.18 SMTP_SEND: Rcvd: 354 Enter mail, end with "." on a line by itself 29-NOV-1993 21:16:05.18 SMTP_SEND: Sent: Received: by lissy1.lif.de (MX V3.3 VAX) id 6707; Mon, 29 Nov 1993 21:16:01 CET 29-NOV-1993 21:16:05.18 SMTP_SEND: Sent: Date: Mon, 29 Nov 1993 21:15:57 CET 29-NOV-1993 21:16:05.19 SMTP_SEND: Sent: From: Hans-Joachim Koch <sys_hjk@lissy1.lif.de> 29-NOV-1993 21:16:05.21 SMTP_SEND: Sent: To: mirror@germany.eu.net 29-NOV-1993 21:16:05.21 SMTP_SEND: Sent: CC: sys_hjk@lissy1.lif.de 29-NOV-1993 21:16:05.21 SMTP_SEND: Sent: Message-ID: <00976485.654110E0.6707@lissy1.lif.de> 29-NOV-1993 21:16:05.21 SMTP_SEND: Sent: Subject: sys_hjk@lissy1.lif.de 29-NOV-1993 21:16:05.21 SMTP_SEND: Sent: 29-NOV-1993 21:16:05.22 SMTP_SEND: Sent: -- 29-NOV-1993 21:16:05.22 SMTP_SEND: Sent: Hans-Joachim Koch, Computer department of Lahmeyer International 29-NOV-1993 21:16:05.22 SMTP_SEND: Sent: Lyoner Strasse 22, D-60528 Frankfurt, Germany 29-NOV-1993 21:16:05.22 SMTP_SEND: Sent: Phone: +49 (69) 6677-642, Fax: +49 (69) 6677-388, Tx: 413478 li d 29-NOV-1993 21:16:05.27 SMTP_SEND: Sent: . 29-NOV-1993 21:16:05.27 SMTP_SEND: will wait 00:10:30.00 for reply. 29-NOV-1993 21:16:05.46 SMTP_SEND: Rcvd: 250 Mail accepted 29-NOV-1993 21:16:05.46 SMTP_SEND: Sent: QUIT 29-NOV-1993 21:16:05.46 SMTP_SEND: Rcvd: 221 lif.de delivering mail 29-NOV-1993 21:16:05.75 Recipient status=00000001 for <mirror@germany.eu.net> 29-NOV-1993 21:16:05.85 Entry now completely processed, no retries needed. 29-NOV-1993 21:16:05.89 *** End of processing pass *** ---------------------------------------------------------------------------- Looks really goods. Always the right address sys_hjk@lissy1.lif.de. But now watch, what comes back from the mail mirror: ---------------------------------------------------------------------------- Return-Path: <daemon@Germany.EU.net> Received: from lif.de (sunny.lif.de) by lissy1.lif.de (MX V3.3 VAX) with SMTP; Mon, 29 Nov 1993 21:16:10 CET Received: from mail.Germany.EU.net by lif.de (4.1/SMI-4.1) id AA08937; Mon, 29 Nov 93 21:20:47 +0100 Date: Mon, 29 Nov 1993 21:17:49 +0100 Message-ID: <199311292017.cM19159@mail.Germany.EU.net> Received: by mail.Germany.EU.net(EUnetD-2.3.0.j) via EUnet id cM19159; Mon, 29 Nov 1993 21:17:49 +0100 Return-Path: nobody From: nobody@Germany.EU.net Subject: Answer from mirror Apparently-To: sys_hjk@lissy1.lif.de Here is the exact copy of the mail you have sent as it has been received by the mirror program on our host. Please check if all the addresses in the fields From, Reply-To, Received-by are fully qualified. Please do NOT use your mailer's 'answer' or 'reply' command to reply to this mail; it won't work because you must specify your e-mail address in the Subject: field. --------------------------------------------------------------- Received: by mail.Germany.EU.net(EUnetD-2.3.0.j) via EUnet id cM19153; Mon, 29 Nov 1993 21:17:48 +0100 Received: from lissy1.lif.de by lif.de (4.1/SMI-4.1) id AA08934; Mon, 29 Nov 93 21:20:44 +0100 Received: by lissy1.lif.de (MX V3.3 VAX) id 6707; Mon, 29 Nov 1993 21:16:01 CET Date: Mon, 29 Nov 1993 21:15:57 CET From: Hans-Joachim Koch <sys_hjk@lifra.lif.de> To: mirror@germany.eu.net Cc: sys_hjk@lifra.lif.de Message-Id: <00976485.654110E0.6707@lissy1.lif.de> Subject: sys_hjk@lissy1.lif.de -- Hans-Joachim Koch, Computer department of Lahmeyer International Lyoner Strasse 22, D-60528 Frankfurt, Germany Phone: +49 (69) 6677-642, Fax: +49 (69) 6677-388, Tx: 413478 li d ---------------------------------------------------------------------------- It's the SAME mail! While sending to mirror@germany.eu.net, my From: address is ok, but in the returned mail, its wrong again. I have NO idea for the reason, never saw something like this before... Very puzzled... Hans. -- Hans-Joachim Koch, Computer department of Lahmeyer International Lyoner Strasse 22, D-60528 Frankfurt, Germany Phone: +49 (69) 6677-642, Fax: +49 (69) 6677-388, Tx: 413478 li d