How to cleanup the cluster lock disk when it gets corrupted

[Hanry Zhou]

Is it possible that the custer lock disk gets corrupted although it is available to access and no problems with the connection?
If there is such situation, what is the way to cleanup the lock disk?

[Ivan Krastev]

What is your SG version ?
In the latest 11.18 you can change lock disk while the cluster is running - http://docs.hp.com/en/B3936-90122/ch08s03.html#cegjbiej

regards,
ivan

[Avinash20]

Hanry : "Is it possible that the custer lock disk gets corrupted although it is available to access and no problems with the connection? "

Do you mean that the Cluster lock disk is claimed in ioscan and works fine, but you are finding some error with regards to CLuster lock in syslog.log ??

Please provide more information ..

There is an cminitlock script which intilizes the cluster lock disk.

Before please let us know you exact requirement?

[Hanry Zhou]

I am not so clean on what is going on, or the detail. But, I heard the entire cluster is down, and two nodes among a 3-nodes cluster were not able to bring up.

It seems like the lock disk (quorum) got corrupted.
when I run cminitlock, should I turn down the cluster, or I can leave it up?

Thanks,

[Avinash20]

"I heard the entire cluster is down

Should I turn down the cluster"

usage: cminitlock [-v] [-t] vg_name pv_name
-t Test the cluster lock only.
-v Verbose output.

This command will initialize a cluster.
lock disk and then query the disk to
validate the disk was initialize
successfully. If the -t option is specified,
the cluster lock is only queried.


Also could you post the syslog /netmft logs of all the three servers

[sujit kumar singh]

hi Hanry,


the cluster might not be coming up as the Clusrer Lock wont be accessible by the nodes of the cluster.

1)Get the Cluster down.

prevent the auto start clluster by editing the /etc/rc.config.d/cmcluster on all the nodes.

change
AUTOSTART_CLCLD=0

2)please replace the cluster lock disk physically that is the first thing reqd. if that is in LVM follow the Disk Replacemnet prcedures.

#vgcfgrestore -n /dev/

after succesful replacement of the disk.
3)vi the cluster ASCII file to ensure that the entry pf te Clutser_Lock_PV and Cluster_Lock_VG are correct.

4)
#cmcheckconf -v -C /etc/
when all seems to be OK and no warnings or error.

If errors then try to rectify them.

then,
5)
#cmapplyconf -v -C /etc/

then start the Cluster (#cmruncl -v)and check all packages by starting the packages also.


6)Once all is OK , shutdown all the application and packages and revert the changes to the /etc/rc.config.d/cmcluster on all the nodes to original.



Start the cluster and packages.


Regards
Sujit

[sujit kumar singh]

Hi

there had been lot of typo in the earlier post by me .
_________________________________________
change
AUTOSTART_CLCLD=0

2)please replace the cluster lock disk physically that is the first thing reqd. if that is in LVM follow the Disk Replacemnet prcedures.

#vgcfgrestore -n /dev/

after succesful replacement of the disk.
____________________________________________


should read actually:

change
AUTOSTART_CMCLD=0

2)please replace the cluster lock disk physically that is the first thing reqd. if that is in LVM follow the Disk Replacemnet procedures.

#vgcfgrestore -n /dev/

after succesful replacement of the disk.


Regards
Sujit