MS Failover cluster issue after converted from Physical RDM to VMDK

[vmnomad]

Hi,

Last week I had to migrate 7 x 2-nodes virtual MS Failover clusters to a new datacenter. All clusters were configured with CAB configuration and all shared disks were in Physical RDM mode.

The target datacenter doesn't provide RDM disks therefore I had to convert all RDM disks into VMDK files and use CIB config prior the migration.

I used vmkfstools to convert RDM into eagdrzeroedthick VMDK files.

5 clusters were migrated successfully, but two cluster got strange issue even though the procedure for all clusters was identical.

So in two clusters I have a problem where the second node in the cluster cannot bring online some of the shared disks. The first nodes in the clusters have no issues with shared disks/services.

But when I try to failover all services to the second node it cannot bring online 1 or 2 shared disks, though it can bring online other shared disks.

VMKernel.log doesn't have any errors or warnings. I have double-checked the configuration of virtual machines and it all looks correct.

Any ideas what else I could check?

[vmnomad · Accepted solution]

Just in case someone else faces the same issue.

I couldn't find the root cause of the issue even though it is definitely the conversion that triggered the problem. However, we found quite a simple workaround - added new shared VMDK files, copied files over and replaced the faulty disks.