VxVM vxdisksetup ERROR V-5-2-43 c6t0d4: Invalid disk device for vxdisksetup

[Nick M_1]

Dear All

I have a problem with some devices on a VSF for RAC clustered system. The cluster is a failback site in a SRDF/A environment. We initially had a problem trying to import Veritas disk groups into this cluster on the failback site, but on further investigation the problem appears to be at device level.

The cluster is Veritas Storage Foundation for RAC 4.1 and operating system is HP-UX 11.23 (Itanium). Powerpath 4.3.1 is in use for multipathing

When we investigated the status of the devices after the disk group import failed, some of these disk groups appeared as being in error on vxdisk list

I called Veritas and they assisted me in running commands to check for registered keys and release any keys on the devices concerned. This did not resolve the matter.

They suggested a reboot and after this all disks appeared on vxdisk list as being in error

I have set scsi3pr on all the cordinator and data devices.

Veritas state this suggests a storage issue.
They said “The vxdisksetup command should complete, as it simply re-writes the public/private region contents. We believe some flag or lock file is preventing us for writing to the device”.

Has anyone seen this "VxVM vxdisksetup ERROR V-5-2-43 c6t0d4: Invalid disk device for vxdisksetup" message before and know how to fix it?

Please see CLI outputs below

Many Thanks

Nick

nick.mander@bt.com


hwneodp1 #
hwneodp1 # /etc/vx/bin/vxdisksetup -i c6t0d4

VxVM vxdisksetup ERROR V-5-2-43 c6t0d4: Invalid disk device for vxdisksetup
hwneodp1 #



hwneodp1 # vxdisk -o alldgs list

DEVICE TYPE DISK GROUP STATUS

c2t0d0 auto:none - - online invalid

c2t1d0s2 auto:LVM - - LVM

c6t0d4 auto - - error

c6t0d5 auto - - error

c6t0d6 auto - - error

c6t1d4 auto - - error

c6t1d5 auto - - error

c6t1d6 auto - - error

c6t2d3 auto - - error

c6t2d4 auto - - error

c7t1d1 auto - - error

c7t1d2 auto - - error

c7t1d3 auto - - error

hwneodp1 #

[Albert_31]

Are you able to talk to the disks on the OS level. a simple dd command would help

dd if=/dev/dsk/ of=/dev/null bs=1024k count=20

The above is a read test and lets you know if the device is accessible or not..

[Nick M_1]

Albert

As per your email I can read the disks -

hwneodp1 # dd if=/dev/dsk/c6t0d4 of=/dev/null bs=1024k count=20
20+0 records in
20+0 records out
hwneodp1 #

hwneodp2 # dd if=/dev/dsk/c6t0d4 of=/dev/null bs=1024k count=20
20+0 records in
20+0 records out
hwneodp2 #

The physical writes to the LUNs seem to be disabled. It seems the LUNs are write protected, but I can't tell whether it's a Veritas or EMC problem at present

Regards

Nick

[Albert_31]

sorry nick,

forgot to put the "Send email to me whenever someone replies " ...hence did not get ur message.

# vxdisk

check if it gives any hints..try the vxprivutil command as well..

[Nick M_1]

Albert

It was the Veritas licence key that expired.I have an emc background - not a Veritas cluster expert. None of the diagnostic messages mentioned an expired
licence key though.

Thanks for your replies

Regards

Nick