Archive-Date: Tue, 9 Jan 2001 09:18:55 -0800 Date: Tue, 9 Jan 2001 12:17:01 -0400 From: feck@fre.fsu.umd.edu Reply-To: MX-List@MadGoat.com To: MX-List@MadGoat.com Message-ID: <009F5DBE.8889DCA0.56@fre.fsu.umd.edu> Subject: Network partner disconnected logical link error Date sent: 9-JAN-2001 12:16:13 We are running MX V5.1-A, VMS 7.1 and UCX 5.0. The following error occurs on ever address that is in one of our listservers. All of the addresses are going to the same site. If I remove all of the addresses and only place a couple addresses to the same site the listserver functions properly. The error message is located below. --Failed delivery to: Address: jdoe@MAIL.FROSTBURG.EDU Status: network partner disconnected logical link Thanks in Advance, Teresa Feck ================================================================================ Archive-Date: Tue, 9 Jan 2001 09:24:21 -0800 Date: Tue, 9 Jan 2001 18:24:12 +0100 From: remi.jolin@telintrans.fr Reply-To: MX-List@MadGoat.com To: mx-list@madgoat.com CC: remi.jolin@telintrans.fr Message-ID: <009F5DF1.D41109CE.28@smtp.tuf.telintrans.fr> Subject: implementing my own rejection policy Is there a way to implement my own rejection policy for the snmp_server ? (a hook for a sharable image ?) Basicaly, what i need is to have a filter based not only on static IP address but in some cases on the IP "name" of the system (some mobile PCs/dynamic DNS). Thanks for help, Remi. ================================================================================ Archive-Date: Tue, 9 Jan 2001 09:41:11 -0800 Message-ID: <008901c07a63$c9ec4730$160610ac@sysmanhome> From: "Ruslan R. Laishev" Reply-To: MX-List@MadGoat.com To: CC: References: <009F5DF1.D41109CE.28@smtp.tuf.telintrans.fr> Subject: Re: implementing my own rejection policy Date: Tue, 9 Jan 2001 20:44:15 +0300 There is a feature provide by ECO-14 (MX 5.1) which can be used for implementation so kind of rejection rules. See an example MX_LOCAL_NET_CHECK.C has been include into this ECO. > Is there a way to implement my own rejection policy for the snmp_server ? > (a hook for a sharable image ?) > > Basicaly, what i need is to have a filter based not only on static IP address > but in some cases on the IP "name" of the system (some mobile PCs/dynamic DNS). > > Thanks for help, > > Remi. > Regards. Mobile:+7 (901) 9713222 http://www.DLS.net - Non-stop VMS-powered ISP in ChicagoLand! http://www.RadiusVMS.com - RADIUS server for OpenVMS project ================================================================================ Archive-Date: Tue, 9 Jan 2001 09:44:08 -0800 Sender: madison@MadGoat.Com Date: Tue, 9 Jan 2001 09:44:04 -0800 From: Matt Madison Reply-To: MX-List@MadGoat.com To: MX-List@MadGoat.com CC: freck@fre.fsu.umd.edu Message-ID: <009F5DA9.2A8AB8C8.1@MadGoat.Com> Subject: RE: Network partner disconnected logical link error MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="iso-8859-1" > We are running MX V5.1-A, VMS 7.1 and UCX 5.0. The following error >occurs on ever address that is in one of our listservers. All of the addresses >are going to the same site. If I remove all of the addresses and only place a >couple addresses to the same site the listserver functions properly. The error >message is located below. > >--Failed delivery to: > > Address: jdoe@MAIL.FROSTBURG.EDU > > Status: network partner disconnected logical link Some SMTP servers use a heuristic for blocking possible spam that causes them to refuse a message if the number of recipients for the message exceeds a certain threshold. Normally, the MX mailing list processor creates a single message for each list posting containing all of the recipients for that posting. You can change that behavior with SET MLF/RECIPIENT_MAXIMUM=n; that will cause MLF to break up a posting into multiple messages, each with no more than the specified number of recipients. (You can also configure this on a per-list basis. You might try setting this to, say, 30 or 40, and see if it helps. -Matt -- Matthew Madison | MadGoat Software | PO Box 556, Santa Cruz, CA 95061 USA madison@madgoat.com http://www.madgoat.com ================================================================================ Archive-Date: Tue, 9 Jan 2001 09:50:51 -0800 Sender: madison@MadGoat.Com Date: Tue, 9 Jan 2001 09:50:47 -0800 From: Matt Madison Reply-To: MX-List@MadGoat.com To: MX-List@MadGoat.com CC: remi.jolin@telintrans.fr Message-ID: <009F5DAA.1B061658.5@MadGoat.Com> Subject: RE: implementing my own rejection policy MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="iso-8859-1" >Is there a way to implement my own rejection policy for the snmp_server ? >(a hook for a sharable image ?) > >Basicaly, what i need is to have a filter based not only on static IP address >but in some cases on the IP "name" of the system (some mobile PCs/dynamic DNS). As already mentioned, the LOCAL_NET_CHECK.C callout in the examples might help here. Also note that V5.2 will support the SMTP authentication extension, which should work better when you have lots of clients with random, dynamic addresses (provided the mail clients also support this extension). -Matt -- Matthew Madison | MadGoat Software | PO Box 556, Santa Cruz, CA 95061 USA madison@madgoat.com http://www.madgoat.com ================================================================================ Archive-Date: Tue, 9 Jan 2001 09:50:58 -0800 Date: Tue, 9 Jan 2001 18:50:47 +0100 From: remi.jolin@telintrans.fr Reply-To: MX-List@MadGoat.com To: laishev@mail.dls.net CC: MX-List@MadGoat.com, remi.jolin@telintrans.fr Message-ID: <009F5DF5.8AD5DCFD.69@smtp.tuf.telintrans.fr> Subject: Re: implementing my own rejection policy The problem with this "plugin" is that it only decides wether the address is inside or outside the network and drives the anti-relay feature according to that. It cannot prevent a PC (only known by its name) to send ANY mail to that gateway. What i need looks more like rejection rules based on the address (add rejection/address=...) than on the "inside_network_address" definition. ----------------------------------------------------------- > There is a feature provide by ECO-14 (MX 5.1) which can be used for >implementation so kind of rejection rules. > > See an example MX_LOCAL_NET_CHECK.C has been include into this ECO. > >> Is there a way to implement my own rejection policy for the snmp_server ? >> (a hook for a sharable image ?) >> >> Basicaly, what i need is to have a filter based not only on static IP >address >> but in some cases on the IP "name" of the system (some mobile PCs/dynamic >DNS). >> >> Thanks for help, >> >> Remi. >> >Regards. >Mobile:+7 (901) 9713222 >http://www.DLS.net - Non-stop VMS-powered ISP in ChicagoLand! >http://www.RadiusVMS.com - RADIUS server for OpenVMS project > ================================================================================ Archive-Date: Tue, 9 Jan 2001 10:41:26 -0800 Date: Tue, 9 Jan 2001 13:39:33 -0400 From: feck@fre.fsu.umd.edu Reply-To: MX-List@MadGoat.com To: MX-List@MadGoat.com Message-ID: <009F5DCA.1002F300.51@fre.fsu.umd.edu> Subject: RE: Network partner disconnected logical link error Date sent: 9-JAN-2001 13:38:05 Thanks Matt for the answer to my problem. Teresa -----Original Message----- From: madison@MadGoat.Com [mailto:madison@MadGoat.Com] Sent: Tuesday, January 09, 2001 12:44 PM To: MX-List@MadGoat.com Cc: freck@fre.fsu.umd.edu Subject: RE: Network partner disconnected logical link error > We are running MX V5.1-A, VMS 7.1 and UCX 5.0. The following error >occurs on ever address that is in one of our listservers. All of the addresses >are going to the same site. If I remove all of the addresses and only place a >couple addresses to the same site the listserver functions properly. The error >message is located below. > >--Failed delivery to: > > Address: jdoe@MAIL.FROSTBURG.EDU > > Status: network partner disconnected logical link Some SMTP servers use a heuristic for blocking possible spam that causes them to refuse a message if the number of recipients for the message exceeds a certain threshold. Normally, the MX mailing list processor creates a single message for each list posting containing all of the recipients for that posting. You can change that behavior with SET MLF/RECIPIENT_MAXIMUM=n; that will cause MLF to break up a posting into multiple messages, each with no more than the specified number of recipients. (You can also configure this on a per-list basis. You might try setting this to, say, 30 or 40, and see if it helps. -Matt -- Matthew Madison | MadGoat Software | PO Box 556, Santa Cruz, CA 95061 USA madison@madgoat.com http://www.madgoat.com ================================================================================ Archive-Date: Wed, 10 Jan 2001 02:28:50 -0800 Date: Wed, 10 Jan 2001 11:28:40 +0100 From: remi.jolin@telintrans.fr Reply-To: MX-List@MadGoat.com To: madison@MadGoat.Com CC: mx-list@madgoat.com, remi.jolin@telintrans.fr Message-ID: <009F5E80.F23E2051.119@smtp.tuf.telintrans.fr> Subject: RE: implementing my own rejection policy Matt, Thanks for the answer, but as i mentioned earlier, the inside/outside check is not enough... a "discard" option would be great. Or the ability to have callouts for all the different rejection checks (address, from, to...) The SMTP authentication is interesting but I've checked the different smtp clients on the "PCs" and some of them don't support it. Regards, Remi. ---------------------------------------------------------------------- From: MI%"madison@MadGoat.Com" (WEB01) 09/01/2001 18:51 09/01/2001 18:51 To: MX-List@MadGoat.com CC: remi.jolin@telintrans.fr Subj: RE: implementing my own rejection policy >Is there a way to implement my own rejection policy for the snmp_server ? >(a hook for a sharable image ?) > >Basicaly, what i need is to have a filter based not only on static IP address >but in some cases on the IP "name" of the system (some mobile PCs/dynamic DNS). As already mentioned, the LOCAL_NET_CHECK.C callout in the examples might help here. Also note that V5.2 will support the SMTP authentication extension, which should work better when you have lots of clients with random, dynamic addresses (provided the mail clients also support this extension). -Matt -- Matthew Madison | MadGoat Software | PO Box 556, Santa Cruz, CA 95061 USA madison@madgoat.com http://www.madgoat.com ================================================================================ Archive-Date: Wed, 10 Jan 2001 06:47:37 -0800 Message-ID: <4FA072E6B8B5D4119AE100508BAED50801F589@scmail> From: Joyce Cogar Reply-To: MX-List@MadGoat.com To: "'MX-LIST@MADGOAT.COM'" Subject: mail fails - see message that queue is full Date: Wed, 10 Jan 2001 09:41:12 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi, one of our remote sites uses vms vax to send mail to mx%"user@domain" and gets %MAIL-E-SENDERR, error sending to user lward@frostburg.edu at = X5(L=D2=E0=F2TPU$ DICUPD ERR=DA=E0=F2 there is more and then messages ends with=20 TPU$ OPENDIC=E2=E0=F2TPU$ STRTOOLNG=F8O*z=EA=E0=F2 TPU$ CNVERR=F2=E0=F2TPU$_BUILTININV?=A49=FA=E0=F2TPU$_REQU IRESDECW -FLQ-E-QUEFULL, MX message queue full I tried stopping mx on both nodes in cluster and creating a new queue then restarting mail but same problem occurs. Any help = appreciated thanks Joyce Cogar Please note new phone numbers Joyce Cogar USM Service Center jcogar@usmsc.edu P.O. Box 530 Voice:301.809.3163 Bowie, Maryland 20718-0530 Fax: 301.809.3130 www.usmsc.edu ================================================================================ Archive-Date: Wed, 10 Jan 2001 06:59:09 -0800 Message-ID: <000b01c07b15$fbd27fa0$859ab1d0@MadGoat.Com> From: "Matt Madison" Reply-To: MX-List@MadGoat.com To: CC: References: <4FA072E6B8B5D4119AE100508BAED50801F589@scmail> Subject: Re: mail fails - see message that queue is full Date: Wed, 10 Jan 2001 06:59:51 -0800 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit >-FLQ-E-QUEFULL, MX message queue full > > >I tried stopping mx on both nodes in cluster and creating a new >queue then restarting mail but same problem occurs. Any help appreciated This generally means that the disk where the message queue resides is getting too full (a clearer error message will be used in MX V5.2). Check the logical name MX_FLQ_DISK_FREE_RESERVED; that value represents a percentage of the total disk space where MX_FLQ_DIR: resides. If the disk has less than that percentage free, MX will refuse to accept messages until the available free space goes above that threshold. If MX_FLQ_DIR: sits on a very large disk, the default value for that logical name (10 percent) may be too large; you can alter the setting by executing MX_DIR:MXCONFIG.COM (option 1 in the menu), then restarting MX. You may also want to check on why the disk is getting full. -Matt ================================================================================ Archive-Date: Thu, 18 Jan 2001 00:53:45 -0800 Date: Thu, 18 Jan 2001 09:53:33 +0100 From: remi.jolin@telintrans.fr Reply-To: MX-List@MadGoat.com To: mx-list@madgoat.com CC: remi.jolin@telintrans.fr Message-ID: <009F64BC.FBDB1210.54@smtp.tuf.telintrans.fr> Subject: mcp reset As MCP STAT shows all the MX processes clusterwide, I think that it would be more consistent to have the MCP RESET command also work clusterwide as the default... Are there many cases where a useless MCP RESET can cause damages ? I know some cases (like modifying a path/rewrite rule...) where a missing reset (you reset only on local node, not on the all of them) can cause many problems... So /cluster should be the default (we are so used to considering a cluster as a single system...). Matt, can you think about it for V5.2 ? Regards, Remi. ================================================================================ Archive-Date: Fri, 26 Jan 2001 14:10:25 -0800 Date: Fri, 26 Jan 2001 17:10:18 -0500 From: "Scott C. Jensen" Reply-To: MX-List@MadGoat.com To: MX-LIST@MADGOAT.COM CC: jensen@ILP.MIT.EDU Message-ID: <009F6B43.521DAAF6.21@ILP.MIT.EDU> Subject: Problem receiving mail from one person Hi folks, I'm hoping someone can help me diagnose a problem we're having on our system. Apparently, there's a small subsidiary office of the company Merck that can't send email to addresses on our system (ilp.mit.edu, VAX/VMS 5.5-2, w/ MX 5.1). We have no problem receiving mail from other Merck offices (or from anyone else, for that matter). Here's a portion of the response a user at the Merck office gets when they send a message to one of our users (henize@ilp.mit.edu): > >postmaster @ TOPCALL on 05/01/2001 11:39:54 > >To: Martin Pellatt/MLGB/Merck@MLGB >cc: > >Subject: Message delayed (henize@ilp.mit.edu) > >Your message has been delayed and is still awaiting delivery >to the following recipient(s): > >henize@ilp.mit.edu >Message delayed > > >Your message is delayed >Message for domain ilp.mit.edu delayed at merck-ltd.co.uk. >Unable to deliver to domain for 48 hours. >Will continue trying for 48 hours. >No action is required on your part. >Last attempt failed because: >Can't connect to host > >Reporting-MTA: dns; merckvc.merckeurolab.ltd.uk >Received-From-MTA: dns; topcall.merckeurolab.ltd.uk (unverified >[62.189.175.132]) >Arrival-Date: Wed, 3 Jan 2001 10:32:17 +0000 > >Final-Recipient: rfc822; henize@ilp.mit.edu >Action: delayed >Status: 4.4.1 (Persistent transient failure - routing/network: no answer >from host) >Will-Retry-Until: Sun, 7 Jan 2001 11:37:26 +0000 > Here's a piece of the analysis that came from Merck's IS group: > >To cut a long story short, the following information may help you >resolve the problem. > >There is no MX record in DNS for ilp.mit.edu - this will either need >correcting or internal routing established to allow delivery to the >ilp sub-domain. The nameserver bitsy.mit.edu provides MX >information for the mit.edu domain. > >This would appear to show that the Merck gateway is fine - and also >explains why mit.edu mails get through OK. > >I hope this helps. Looks like you need to add a MX record in DNS >for ilp.mit.edu to solve the problem. > This seems wrong to me. There is indeed no MX record for ilp.mit.edu in the DNS system. My (limited) understanding of DNS is that we would only want a DNS record for ilp.mit.edu if we wanted mail sent here to actually go somewhere else. Is that not true? FYI, here's the two DNS records that do cover our host: ANSWER #1: ILP.MIT.EDU., HINFO, INTERNET, 0 06:00:00, CPU=DEC/VAX-6220, OS=VMS ANSWER #2: ILP.MIT.EDU., A, INTERNET, 0 06:00:00, Addr=18.48.0.100 Anyway, my feeling is that the problem doesn't lie on our end, since we've been receiving mail on this system for years with no problems. Please let me know if you have any suggestions about this. Thanks! ---SCJ Scott C. Jensen Manager, Office of Information Services MIT Corporate Relations - Industrial Liaison Program Room E38-576 292 Main Street Cambridge, MA 02139 617/253-0441 FAX: 617/258-0150 Email: jensen@mit.edu ================================================================================ Archive-Date: Sun, 28 Jan 2001 08:27:36 -0800 Sender: madison@MadGoat.Com Date: Sun, 28 Jan 2001 08:27:32 -0800 From: Matt Madison Reply-To: MX-List@MadGoat.com To: MX-List@MadGoat.com CC: jensen@ilp.mit.edu Message-ID: <009F6C8C.9F4EF814.2@MadGoat.Com> Subject: RE: Problem receiving mail from one person MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="iso-8859-1" > I'm hoping someone can help me diagnose a problem we're having on our >system. Apparently, there's a small subsidiary office of the company Merck >that can't send email to addresses on our system (ilp.mit.edu, VAX/VMS >5.5-2, w/ MX 5.1). We have no problem receiving mail from other Merck >offices (or from anyone else, for that matter). Here's a portion of the >response a user at the Merck office gets when they send a message to one of >our users (henize@ilp.mit.edu): [...] >>Final-Recipient: rfc822; henize@ilp.mit.edu >>Action: delayed >>Status: 4.4.1 (Persistent transient failure - routing/network: no answer >>from host) >>Will-Retry-Until: Sun, 7 Jan 2001 11:37:26 +0000 >> > > > Here's a piece of the analysis that came from Merck's IS group: > >> >>To cut a long story short, the following information may help you >>resolve the problem. >> >>There is no MX record in DNS for ilp.mit.edu - this will either need >>correcting or internal routing established to allow delivery to the >>ilp sub-domain. The nameserver bitsy.mit.edu provides MX >>information for the mit.edu domain. >> >>This would appear to show that the Merck gateway is fine - and also >>explains why mit.edu mails get through OK. >> >>I hope this helps. Looks like you need to add a MX record in DNS >>for ilp.mit.edu to solve the problem. >> > > This seems wrong to me. There is indeed no MX record for ilp.mit.edu >in the DNS system. My (limited) understanding of DNS is that we would only >want a DNS record for ilp.mit.edu if we wanted mail sent here to actually >go somewhere else. Is that not true? FYI, here's the two DNS records that >do cover our host: > >ANSWER #1: ILP.MIT.EDU., HINFO, INTERNET, 0 06:00:00, CPU=DEC/VAX-6220, OS=VMS >ANSWER #2: ILP.MIT.EDU., A, INTERNET, 0 06:00:00, Addr=18.48.0.100 > > Anyway, my feeling is that the problem doesn't lie on our end, since >we've been receiving mail on this system for years with no problems. >Please let me know if you have any suggestions about this. You are right, their analysis is incorrect. Per RFC974, if there is no MX RR for a name, the sending SMTP should just try to look up the address and send the message there. -Matt -- Matthew Madison | MadGoat Software | PO Box 556, Santa Cruz, CA 95061 USA madison@madgoat.com http://www.madgoat.com