Hi.

Running OESSP2linux 2Node Cluster with HP EVA4000 as shared storage in a VMWARE ESX 3.01 environment.
This cluster has been in production for a couple of months and I seem to have gotten all the issues with oes and vmware resolved.

Now I'm trying to do an online resizing of a pool which works just fine in our old NW6SP5 cluster.

Scenario1:
Lun1 on SAN is 10GB and presented to the servers as /dev/sdb

Created a 5GB pool called test1
Created a volume that uses all the space in pool test1
Expanded pool test1 to 10GB

I have tried this in both nssmu and iManager and everything works just as expected, no issues whatsoever.

Scenario2
Lun1 on SAN is 5GB and presented to the servers as /dev/sdc

Created a 5GB pool called test2
Created a volume that uses all the space in pool test2
Expanded Lun1 on SAN to 10GB
Used sfdisk -R /dev/sdc on both nodes so they can see the new diskarea

Both in nssmu and in iManager the device is now seen as having 10GB of space

When trying to expand the pool this time in either interface i get very different kinds of errors, sometimes it seems to work but it is
impossible to activate the pool, sometimes nssmu gives me an error "22" and all other operations in nssmu hangs. Sometimes the free space is all
of a sudden several million gigabytes.

I suspect that I need to do something else to get evms to understand the new disk-geometry? I can not find ANYTHING in the oes documentation
about this and in the forums I only found one post from someone that seems to have the exact same issues but no resolution.

A rescan of the scsi bus does nothing when it comes to detecting new disk-geometry.

Any hints?

/anders westerberg