Unable to determine a unique identifier for physical volume

[Rajendra prasad NVR]

we have two node cluster which is running fine. But when i try to get the ascii file gettting below errors and not generating the ascii file. Is there any way to ignore this error and generate ascii file.

baleadm2:/etc/cmcluster#cmgetconf -c zale_clu1
bale_clu1.ascii.2
Unable to determine a unique identifier for physical volume /dev/dsk/c4t2d4 on node zaleadm1. Use pvcreate to give the disk an identifier.
Unable to determine a unique identifier for physical volume /dev/dsk/c4t2d5 on node zaleadm1. Use pvcreate to give the disk an identifier.

Thanks in advance.

[Bhadresh]

Hi,

What version of Serviceguard and HP-UX you are using?

Regards,
Bhadresh

[Vijaykumar_1]

Im getting confused here..

Can you provide the name of the cluster and also the name of the package along with the two hostnames running the cluster..

[Bhadresh]

Hi,

Look at the Logical Volume Manager information in all cluster nodes whether it is in sync or not, if not vgimport/vgexport.

Similar thread:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1213268

Regards,
Bhadresh

[Kapil Jha]

which OS, which service guard?

I suppose these disks are not supported by MCSG.

See
http://docs.hp.com/en/B3935-90091/ch01s06.html

and
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=744311

BR,
Kapil+

[njia_1]

Hi there

maybe /etc/lvmtab corrupted, do the following to fix this.
mv /etc/lvmtab /etc/lvmtab.save
vgscan

If it says can not match the disk to a vg then you need to vgimport again

thanks

[johnsonpk]

Hi Rajendra,

Is it 11.23 Serviceguard A.11.17.00 ?

it seems it has the said bug and have a fix also,

apply PHSS_34337 or PHSS_38744 for this issue





Regards!
Johnson

[Rajendra prasad NVR]

Thank you for all who replies to my request.

ServiceGuard A.11.18.00 Serviceguard SD Product

version of os is
HP-UX baleadm2 B.11.23 U 9000/800 198434636 unlimited-user license

Now the cluster is up i am not sure if we do reimporting the vg's will effect the cluster. It was production servers need downtime to do any acitivity.

[rariasn]

Hi,

# cmgetconf -K -c zale_clu1 bale_clu1.ascii.2


-K This option causes the probing of volume groups to be skipped, reducing the time for the command to complete.
This option is not available for Linux. Note that if this option is used, cmgetconf will not add the list of cluster-aware volume groups to the resulting cluster configuration file. You must add the cluster-aware volume groups to the cluster configuration file before running cmapplyconf, or use the -K option of
cmapplyconf when applying the generated configuration file. Failure to do so will result in all volume groups in the cluster being unusable in packages or for the cluster lock.

rgs

[Rajendra prasad NVR]

-K option was not there in HPUNIX.
Any other option available.

[S.N.S]

>>Now the cluster is up i am not sure if we do reimporting the vg's will effect the cluster. It was production servers need downtime to do any acitivity.<<

As per my knowledge - the vgreimport procedure doesn't need downtime --- the vgexport is done with preview mode >> vgimport -p

Check the attachment for details.

HTH
SNS

[S.N.S]

Sorry it is vgexport -p

[Rajendra prasad NVR]

reimporting solved the problem