vgdisplay -v <CLUSTER_LOCK_VG>

[ManojK_1]

Hi,

I have configured Two HP Unix Servers in Cluster for SG NFS tool kit.

After starting the cluster, if i am trying to check the cluster lock vg status it is showing the following output from both the cluster nodes.

#vgdisplay -v vgsgq
vgdisplay: Volume group not activated.
vgdisplay: Cannot display volume group "vgsgq".

Is it normal? or the cluster lock VG should be activated in exclusive mode in any of the server where NFS package active.

[sujit kumar singh]

Hi

This is not necesary, if the cluster lock VG is not a part of the package that would activate that this wont be activated or you shall be able to see that as activated.
If the cluster lock VG is also part of the package VGs then this can be seen as activated as the package shall start.

in case you wish to activate this VG you can do that by using vgchange -a e VG_Name on one of the niodes and then shall be able to see that in vgdisplay.

the cluster lock VG can remain deactivated if that is not the part of the Package vGs.
regards
sujit

[Michael Steele_2]

Hi

a) Is the lock disk a lun or a local disk?
b) Are you using vxvm?
c) What does this command do?

cmdisklock reset /dev/dsk/c0t1d1

d) What are the errors in syslog.log?
e) Why can't you activate vgsgq from the command line with the vgchange -a y -c y vg command. Note: Check the man page for correct arguements.

[sujit kumar singh]

Hi


if the cluster lock VG is Cluster aware that, then this can be activated on any one of the nodes in an exclusive mode by vgchange -a e and then on that node you can run vgdisplay -v to have a look at that.

The VG needs to be cluster aware, this can be done by putting that VG_Name in the cluster config ASCII file or when the cluster is UP by running vgchange -a n then vgchange -c y , then when Cluster aware flag is set you can activate the VG on one of the nodes in exlusive mode using the -e option of the vgchange command, and then can use vgdisplay.


can you post the cluster config ascii file here.
regards
sujit

[ManojK_1]

Hi,

Thanks for the reply.

From the replay what i understood that the cluster will not make the cluster lock vg active if the vg is not part of any package in the cluster.

I am not using vxvm and the lun is presented from EVA.

#cmdisklock reset /dev/disk/disk14
WARNING: Cluster lock LUN /dev/disk/disk14 is used by LVM: . Until this situation is corrected, a single failure could cause all nodes in the cluster to crash.
To force use of cluster lock LUN on this disk, destroying all other data on the disk, run 'cmdisklock -f reset /dev/disk/disk14' after ensuring that all active nodes in the cluster have logged this message.
/dev/disk/disk14 is inaccessible
Resetting cluster lock device /dev/disk/disk14
Cluster lock reset failed
error


There is no error in syslog.

[Michael Steele_2]

Hi

a) Run

cmdisklock reset /dev/disk/disk14

...from both nodes. Verify the same message posted from both nodes.

b) Are you storing data on the lock lun? If so, don't. That's your problem.

Keep the following points in mind when choosing a LUN for a lock

LUN:
â ¢ All the cluster nodes must be physically connected to the lock LUN.

â ¢ A lock LUN must be a block device.

â ¢ All existing data on the LUN will be destroyed when you configure it
as a lock LUN.

â ¢ A lock LUN cannot also be used in an LVM physical volume or VxVM
or CVM disk group.

. A lock LUN cannot be shared by more than one cluster.

â ¢ A lock LUN cannot be used in a dual-lock configuration.

â ¢ You do not need to back up the lock LUN data, and in fact there is no
way to do so.
A lock LUN needs only a small amount of storage, about 100 KB.

â ¢ If you are using a disk array, create the smallest LUN the array will
allow,

â ¢ If you are using individual disks, use either a small disk, or a portion
of a disk. On an HP Integrity server, you can partition a disk;