hwmgr -view devices output after dsfmgr -m

From: David J. DeWolfe (sxdjd@ts.sois.alaska.edu)
Date: Mon Aug 15 2005 - 14:26:57 EDT


All;

Recently I created a couple of new vdisks on our EVA5000 and presented them
to our 4 node GS1280 cluster running Tru64 5.1b PK4. I used hwmgr to
refresh the components and then saw the 2 new devices in the "hwmgr -v d"
output. All looked as it should. I then renamed the device special files
via "dsfmgr -m <old device> <new device>. The normal messages appeared on
the screen and the command completed successfully. The <old device> files
no longer exist in /dev/disk and the <new device> files exist in /dev/disk.
However, "hwmgr -v d" does not report the correct info. Here's a snippet of
the output from the "hwmgr -v d" command:

> 1345: /dev/disk/dsk12717c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12717
> 1865: /dev/disk/dsk12801c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12801
> 1876: /dev/disk/dsk12900c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12900
> 2113: /dev/disk/dsk12720c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12720
> 2114: /dev/disk/dsk12721c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12721
> 3249: dsk22410 COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12722
> 3250: dsk22411 COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12723
> 3251: /dev/disk/dsk22412c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12582
> 3252: /dev/disk/dsk22413c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12583

the 2 problem disks are dsk22410 and dsk22411. When I first created and
presented the disks they looked like all the others above in that the
output showed /dev/disk/dsk22410 and /dev/disk/dsk22411. In other words, it
looked like:

> 3249: /dev/disk/dsk22410c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12722
> 3250: /dev/disk/dsk22411c COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12723

Then I issued the following 2 commands:

         # dsfmgr -m dsk22410 dsk12722
         # dsfmgr -m dsk22411 dsk12723

after which the "hwmgr -v d" output changed to:

> 3249: dsk22410 COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12722
> 3250: dsk22411 COMPAQ HSV110 (C)COMPAQ IDENTIFIER=12723

I'm guessing that perhaps the hwmgr database is confused. Would verifying
the dsfmgr database (dsfmgr -v) be worthwhile or applicable? I presume that
a reboot would resolve this issue but was wondering if there was some way
short of that to get hwmgr to reflect the correct info.

TIA.

David
mailto:sxdjd@ts.sois.alaska.edu



This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:50:22 EDT