Lock disk creation documentation

[Mike Smith_33]

I have moved the data my existing 2 node, 11i v1 Serviceguard cluster to a remote datacenter. I am having problems getting the cluster to form. Cmcheckconf has been core dumping on me. I contacted HP and they pointed towards a lock disk problem, sent me the doc and closed the ticket.

I can not find anywhere in the documentation the exact steps for creating the lock disk. I have created and populated all the other disks using pvcreate, lvm and vg commands. Hardware is identical to the existing system even down to the i/o paths. I have no problem reading the manual, I just haven't been able to find specifics.

[Kenan Erdey]

Hi,

check lock pv name in cluster ascii config file. and do cmapplyconf again.

Kenan.

[Steven E. Protter]

Shalom,

Perhaps link in the doc you are using, because its on my paper copy.

As noted above, change the config file then cmcheckconf cmapplyconf

SEP

[Mike Smith_33]

First off, thanks for the replies.

I am confused as to your answers since I stated that the hardware paths down to the i/o are identical. The lock pv device name is the same as it was on the system I moved from. Am I missing something?

[Mike Smith_33]

Here is what I am getting.

root@xxxxxx$ cmcheckconf -v -C cluster.ascii

Checking cluster file: cluster.ascii
Checking nodes ... Done
Checking existing configuration ... Done
Gathering configuration information ... Done
Gathering configuration information ... Done
Gathering configuration information ..
Gathering storage information ..
Found 8 devices on node xxxxxx
Found 8 devices on node yyyyyy
Analysis of 16 devices should take approximately 2 seconds
0%----10%----20%----30%----40%----50%----60%----70%----80%----90%----100%
Found 7 volume groups on node xxxxxx
Found 6 volume groups on node yyyyyy
Analysis of 13 volume groups should take approximately 1 seconds
0%----10%----20%----30%----40%----50%----60%----70%----80%----90%----100%
.....
Gathering Network Configuration ......Abort(coredump)

[Kenan Erdey]

Hi,

check this:

https://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1088044

[melvyn burnard · Accepted solution]

Well first of all, you do not need the cluster lock disc to be able to start your cluster, so there is something more serious wrong here.
Secondly, you have to have the shared VG activated on the node on which you are going to do the cmcheckconf/cmapplyconf for it to be marked as a cluster lock disk.

I suspect there is something more substantially wrong here.
I would be inclined to consider taking a backup of teh /etc/cmcluster directoy on each node, try using cmquerycl to recreate a cluster ascii file as a test, then delete the cluster binary using either cmdeleteconf or simply remove the cmclconfig file on each node and see if cmcheckconf/cmapplyconf will now work.
also verify your patch versions of SG are at the very latest.

[Mike Smith_33]

Thanks to everyone that replied, I really do appreciate it. Mel pretty much hit it on the head. The main things I did were to delete all existing, vgexport the main nodes vgs over to a mapfile, import from the other node. Cmquerycl and modify cluster ascii file, check the config, apply the config, run the cluster.

I do still have a package problem but I the main problem is past now.