Original post at the bottom. I've been out of town and just got back to the
problem this afternoon.
>Terse question:
>
>After adding 2 new drives to a scsi chain, the tape drive that was already
>there is no longer available...
>
>Wordy version:
>
>I've just moved a couple of Seagate 18GB drives from a Solaris box over to
>an Alpha box (running 4.0F).. I did what I usually do - shutdown, turn off
>the SCSI stuff, connect the cables and terminator at the end of the chain,
>turn the scsi stuff back on, and turn on the Alpha.
Updated info:
A number of folks had suggestions (thanks ! - I'll list all when I've a
resolution or decide I've failed to resolve the problem) and I've tried
some suggestions and a couple of things of my own.
(1) Shutting all down and booting only with the Seagate drives connected
gives me 2 drives, mounted correctly and no errors.
Apr 25 14:20:41 bilbo vmunix: isp1: Fast RAM timing enabled.
Apr 25 14:20:41 bilbo vmunix: scsi3 at isp1 slot 0
Apr 25 14:20:41 bilbo vmunix: rz28 at scsi3 target 4 lun 0 (LID=2) (SEAGATE
ST118273W 6244) (Wide16)
Apr 25 14:20:41 bilbo vmunix: rz30 at scsi3 target 6 lun 0 (LID=3) (SEAGATE
ST118273W 5698) (Wide16)
(2) Shutting all down and booting only with the tape drive connected (Sony
AIT in ADIC library) results in no problems.
Apr 25 15:41:34 bilbo vmunix: scsi3 at isp1 slot 0
Apr 25 15:41:34 bilbo vmunix: tz25 at scsi3 target 1 lun 0 (LID=2)
(SONY SDX-300C 0400) (Wide16)
Apr 25 15:41:34 bilbo vmunix: lvm0: configured.
(3) Reordering the SCSI chain so that the seagates are in the middle
results in a failure to boot. The drive activity lights come on, and stay
on. Error -
Apr 25 14:35:22 bilbo vmunix: isp1 at pci1000 slot 10
Apr 25 14:35:22 bilbo vmunix: isp1: QLOGIC ISP1040B/V2
Apr 25 14:35:22 bilbo vmunix: isp1: Firmware revision 5.57 (loaded by console)
Apr 25 14:35:23 bilbo vmunix: isp1: Fast RAM timing enabled.
Apr 25 14:35:23 bilbo vmunix: cam_logger: CAM_ERROR packet
Apr 25 14:35:23 bilbo vmunix: cam_logger: bus 3 target 4 lun 0
Apr 25 14:35:23 bilbo vmunix: ss_perform_timeout
Apr 25 14:35:23 bilbo vmunix: timeout on disconnected request
Apr 25 14:35:23 bilbo vmunix: Active CCB at time of error
Apr 25 14:35:23 bilbo vmunix: cam_logger: CAM_ERROR packet
Apr 25 14:35:23 bilbo vmunix: cam_logger: bus 3
Apr 25 14:35:23 bilbo vmunix: isp_mailboxcomplete
Apr 25 14:35:23 bilbo vmunix: Timeout on mailbox command completion -
scheduling chip reinit
Turning the seagate drive box off results in completion of booting.
During all of this switching, I am moving the SCSI terminator to the
correct location.
A couple of folks suggested that I might be mixing single & differential
SCSI on a single chain. I don't think so - the mix of Sony AIT and Seagate
drives worked fine on the solaris box. Same seagates, different AIT drive.
But in any case, how do I find out if I'm mixing single & differential SCSI
on the same chain?
In answer to other suggestions/comments:
(1) The devices are all there. They do not need to be made.
(2) Device numbers haven't changed - particularly the tape.
(3) The bus/cable length is not excessive - there is a 12" and a
36" cable (only) on the chain, and when the seagate drives are turned off
and in the middle, the tape drive still works.
(4) Several folks suggested I rebuild the kernel. Since all
devices work separately, why? The drivers must be there and the special
devices definitely are there already.
Any further suggestions before I bite the bullet and see if I can get an
answer out of Compaq/Dec technical support?
What I'm really trying to accomplish here is avoidance of the excessive
charges for buying cassette mounted drives to plug into an ES40 chassis and
NFS mount these 18G drives from a 433au sitting on my desk. I just have to
get the drives and the tape system working simultaneously.....
At 03:33 PM 4/12/2000 -0500, Chad Price wrote:
>Terse question:
>
>After adding 2 new drives to a scsi chain, the tape drive that was already
>there is no longer available...
>
>Wordy version:
>
>I've just moved a couple of Seagate 18GB drives from a Solaris box over to
>an Alpha box (running 4.0F).. I did what I usually do - shutdown, turn off
>the SCSI stuff, connect the cables and terminator at the end of the chain,
>turn the scsi stuff back on, and turn on the Alpha.
>
>Well, everything gets recognized on reboot - Note the last 3 lines showing
>the tape and the drives.
>
>Apr 12 10:41:28 bilbo vmunix: isp0 at pci1000 slot 4
>Apr 12 10:41:28 bilbo vmunix: isp0: QLOGIC ISP1040B/V2
>Apr 12 10:41:28 bilbo vmunix: isp0: Firmware revision 5.57 (loaded by console)
>Apr 12 10:41:28 bilbo vmunix: scsi2 at isp0 slot 0
>Apr 12 10:41:28 bilbo vmunix: rz16 at scsi2 target 0 lun 0 (LID=1)
>(COMPAQ BB00921B91 3B05) (Wide16)
>Apr 12 10:41:28 bilbo vmunix: isp1 at pci1000 slot 10
>Apr 12 10:41:28 bilbo vmunix: isp1: QLOGIC ISP1040B/V2
>Apr 12 10:41:28 bilbo vmunix: isp1: Firmware revision 5.57 (loaded by console)
>Apr 12 10:41:28 bilbo vmunix: isp1: Fast RAM timing enabled.
>Apr 12 10:41:28 bilbo vmunix: scsi3 at isp1 slot 0
>Apr 12 10:41:29 bilbo vmunix: tz25 at scsi3 target 1 lun 0 (LID=2)
>(SONY SDX-300C 0400) (Wide16)
>Apr 12 10:41:29 bilbo vmunix: rz28 at scsi3 target 4 lun 0 (LID=3)
>(SEAGATE ST118273W 6244) (Wide16)
>Apr 12 10:41:29 bilbo vmunix: rz30 at scsi3 target 6 lun 0 (LID=4)
>(SEAGATE ST118273W 5698) (Wide16)
>
>But when I'm in multi-user mode, the tape drive (SONY SDX-300C) is invisible -
> "mt -f /dev/nrmt2h status"
>fails to find the tape device. The drives work fine - formatted and
>mounted as ufs.
>
>Shut it all down, disconnect the drives and put the terminator in the
>right place for the chain, and reboot - now I have a working tape drive.
>
>The only thing I saw in the archives was June 1996 and seemed to be
>overkill and a much earlier version of DU - I don't think I need to make a
>new kernel - particularly as everything is (at one point or another)
>recognized...
>http://www.ornl.gov/its/archives/mailing-lists/tru64-unix-managers/1996/06/m
>sg00462.html
>
>Anyone got any ideas? I obviously want both the drives and the tape to be
>active simultaneously...
>
>Chad
>
>Chad Price
>Systems Manager, Genetic Sequence Analysis Facility
>University of Nebraska Medical Center
>986495 Nebraska Medical Center
>Omaha, NE 68506-6495
>cprice_at_molbio.unmc.edu
>(402) 559-9527
>(402) 559-4077 (FAX)
>
Chad Price
Systems Manager, Genetic Sequence Analysis Facility
University of Nebraska Medical Center
986495 Nebraska Medical Center
Omaha, NE 68506-6495
cprice_at_molbio.unmc.edu
(402) 559-9527
(402) 559-4077 (FAX)
Received on Tue Apr 25 2000 - 21:59:02 NZST