[javsc]

Hi everyone,
After extending a concurrent VG in a HACMP resource group and mirroring the VG's data onto the new disk. The output of an lspv on the standby node looks like:
hdisk7 <PVID> oraclevg
hdisk8 <PVID> oraclevg

This would be fine except that prior to extending/mirroring the standby node looked like this:
hdisk7 <PVID> oraclevg (concurrent)

Is there a way I can get the VG's volumes to list as concurrent without shutting down HACMP? Will keeping it as-is affect the failover process? The disks do list as concurrent on the active node.

Thanks!

[ogniemi]

I don't know if I clearly understand you but...

Prior to extend a VG by a new disk in a cluster I was always discovering this disk on one node, running "chpv -l hdiskX -a pv=yes" to assign a PVID to the disk, and then discover this disk on second cluster node. Starting from this point I had the disk on both nodes with the same PVID but not accigned to any VG yet.

Then I went to CSPOC utility and extended a VG by this new disk and after that "lspv" command run on both cluster nodes showed me that VG has a new member disk. All other operations like mirroring I was also doing from CSPOC and all ODM informations were kept synced / up-to-date on both nodes.

[mansoorkalim]

Hi,

Did you mean that running chdev -l hdiskx -a pv=yes on the node which has the resources, active node.

Thanks

[halley]

Hi,
You will need to run verify/ synch again to get it back to the state of concurrent on both.
Regards

[javsc]

Hi, thanks for your replies. The verify/sync was completing without error and not setting the (concurrent) attribute of the standby node's PVs. Actually, we found that gsclvmd kept failing which is apparently required to sync the lvm config. We've upgraded HA to 5.2.8 at IBM support's recommendation and so far we haven't had these problems as we proceed with our project.

Thanks again! =)