can't get hardware address for device

[CGI Unix Admin]

Hi,

I have a vg01 with a dump device named /dev/vg01/lvdump.

I just migrated it from a EMC storage to Hitachi using lvm mirroring. Rebooted twice no issues. Now i installed HDLM for multipathing control and rebooted. now i get the following error.

Configure system crash dumps
Output from "/sbin/rc1.d/S080crashconf start":
----------------------------
/dev/vg01/lvdump: error: can't configure firmware access to device
EXIT CODE: 1


Any ideas? Im running 11.11 on a superdome and the volume resides on a USP600

[CGI Unix Admin]

Almost forgot here is some more info

# grep lvdump /etc/fstab
/dev/vg01/lvdump ... dump defaults 0 0

# lvdisplay -v /dev/vg01/lvdump |more
--- Logical volumes ---
LV Name /dev/vg01/lvdump
VG Name /dev/vg01
LV Permission read/write
LV Status available/syncd
Mirror copies 0
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 16384
Current LE 4096
Allocated PE 4096
Stripes 0
Stripe Size (Kbytes) 0
Bad block NONE
Allocation PVG-strict/contiguous
IO Timeout (Seconds) default

--- Distribution of logical volume ---
PV Name LE on PV PE on PV
/dev/dsk/c13t0d1 4096 4096

--- Logical extents ---
LE PV1 PE1 Status 1
00000 /dev/dsk/c13t0d1 00000 current
00001 /dev/dsk/c13t0d1 00001 current
00002 /dev/dsk/c13t0d1 00002 current
[truncated by me]

[Michael Steele_2]

Hi CGI Unix Admin:

Got a name? CGI Unix Admin sounds like a title of courtesy.

Dump is swap. Lvol2 under vg00. It is contiguous, as are lvol1 and 3. When the O/S loads it starts reading sequentially from the first address spot in lvol1, are right after the boot lif partition. It is address 0000 of the 2nd partition on the boot disk. So it reads from 0000 through lvols 1, 2 and 3 in order to load, or it fails.

You've built your dump in vg01. This is the breaking of a basic rule of HP-UX. Perhaps you mean to make it a repository for saving a dump after a panic. If so then this is not the same. Review these man pages, crashconf, savecrash and lvlnboot -d. Also cross reference by O/S versions. Note: Usually dumps are written into /var/adm/crash during a panic.

Here's a good 11i v3 doc

http://whitepapers.techrepublic.com.com/abstract.aspx?docid=354970

Here are good hp-ux 11.00 doc

http://docs.hp.com/en/951/syscrash.html

http://docs.hp.com/en/B2355-90950/ch05s05.html

[CGI Unix Admin]

Sorry, I'm using the ID they have listed for this account. My name is Richard.

If I compare with other hosts that are configure with same OS and similar layout, i have here

Configure system crash dumps
Output from "/sbin/rc1.d/S080crashconf start":
----------------------------
/dev/vg01/lvdump: warning: no link redundancy, only primary path used for dump
EXIT CODE: 0

[prdap1:/root]# grep dump /etc/fstab
/dev/vg01/lvdump ... dump defaults 0 0

it hasn't been migrated yet but will be soon.

[Michael Steele_2]

Hi Richard:

This looks like a SAN zoning issue. If this were an EMC disk array, instead of an Hitachi UPS600, then I would run powermt and verify my alternate paths, and array id. You'll have to find the equivalent for the UPS600. But, with HP-UX and LVM you can run:

vgdisplay -v vg01 | grep "PV Name"
-and-
pvdisplay -v /dev/dsk/c#t#d# | more
-and-
ioscan -fnC fc (* to check for CLAIMED HBA's *)

Note: The 'fc' name might be a different class name for you.