Conflicting tape devices from 'boot -r'

From: Deb (dbh@tickleme.llnl.gov)
Date: Mon Sep 29 2003 - 12:18:02 EDT


Hi,

New to the list here. Having trouble with Solaris 2.7 creating the
last two of 8 devices for 8 SDLT tape drives. This system has the
latest patch rev, and also has patch 107460 installed.

Scenerio:

        Sun E450
        2 Sun Dual Port Ultra SCSI cards (X6541).
        Quantum/ATL P2000, tape library w/8 SDLT drives.

When fully connected the configuration is:
(Drive #'s are the physical drive number, as seen from the tape library)

        Tape Library E450 Slots
        ===========================================================
        Drives 1 & 2 and P2000 controller Slot 4 Port A
        Drives 3 & 4 Slot 4 Port B
        Drives 5 & 6 Slot 7 Port A
        Drives 7 & 8 Slot 7 Port B
        
(more config data below the problem text)

Problem:
The problem occurs after "boot -r" and the devices are generated (after
/dev/rmt/* has been deleted). While it seems that all the devices are
created, drives 5-8 (ie. /dev/rmt/4* through /dev/rmt/7*) are not accessible
via the 'mt' command.

Next, if I disconnect a cable from slot 7 (either cable), and reboot, the "mt"
command is now successfull in communicating with the 2 remaining drives.

I see that the devices that are created for drives 5-8 are not quite
sequenced in the same way as drives 1-4. But, even if I move the scsi card
to a different slot, they are similarly numbered. See below for how the
config looks. All drives are working, as confirmed by moving the cable to
different SCSI ports to ensure that the drives are accessible.

Perhaps it is naive to think this way, but I would have expected drives 5-8
to come up with devices something like

/pci@6,2000/scsi@1 - Target 5 and 6
/pci@6,2000/scsi@1,1 - Target 7 and 8

But, this never happens. I know that scsi id 7 is supposed to be reserved
for the host (from the E450 manual), so I'm thinking this is perhaps why I'm
not getting Target 7? If this is the reason, how can I force 2 working
device nodes?

So, I'm stuck now with on 6 of eight useful drives.

Where do I go from here???

I hope someone has Been Here, Done This, and can share their resolution. Of
course, I will Summarize to the list.

Thanks,

deb

------------------------------------------------------------------
At the prom, probe-scsi-all shows:

{0} ok probe-scsi-all
/pci@6,2000/scsi@1,1
Target 5
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E
Target 6
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E

/pci@6,2000/scsi@1
Target 5
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E
Target 6
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E

/pci@6,4000/pci@4/scsis@4/mscsi@2,0

/pci@6,4000/pci@4/scsis@4/mscsi@1,0

/pci@6,4000/pci@4/scsis@4/mscsi@0,0

/pci@6,4000/pci@3/scsis@4/mscsi@2,0

/pci@6,4000/pci@3/scsis@4/mscsi@1,0

/pci@6,4000/pci@3/scsis@4/mscsi@0,0

/pci@4,4000/scsi@4,1
Target 3
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E
Target 4
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E

/pci@4,4000/scsi@4
Target 0
  Unit 0 Removable Device type 8 ATL P2000 63182502.2
Target 1
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E
Target 2
  Unit 0 Removable Tape QUANTUM SuperDLT1 2E2E

------------------------------------------------------------------

After boot -r, the devices are created (from a long list of /dev/rmt/*cbn):

/dev/rmt/0cbn -> ../../devices/pci@4,4000/scsi@4/st@1,0:cbn
/dev/rmt/1cbn -> ../../devices/pci@4,4000/scsi@4/st@2,0:cbn
/dev/rmt/2cbn -> ../../devices/pci@4,4000/scsi@4,1/st@3,0:cbn
/dev/rmt/3cbn -> ../../devices/pci@4,4000/scsi@4,1/st@4,0:cbn
/dev/rmt/4cbn -> ../../devices/pci@6,2000/scsi@1/st@5,0:cbn
/dev/rmt/5cbn -> ../../devices/pci@6,2000/scsi@1/st@6,0:cbn
/dev/rmt/6cbn -> ../../devices/pci@6,2000/scsi@1,1/st@5,0:cbn
/dev/rmt/7cbn -> ../../devices/pci@6,2000/scsi@1,1/st@6,0:cbn
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:27:12 EDT