RDM - VMWare - Failover Cluster - Server 2016 - Disk increase / extend / expand

[KMHH]

Hi all

I have a 2-node cluster that comprises of;

• 2 x Windows 2016 servers on VMWare 6.5 connected to HP SAN.

These nodes have an RDM attached.  This RDM is in physical mode.

When extending the RDM on the SAN (which was done whilst the VMs were running) I noticed the following:

• The guest OS immediately sasw the increased size.
• The VM - Edit Settings - shows the original size and not the increased size.

Even after issuing command:

vim-cmd vmsvc/reload 'vm-id'

the size seen by the VM in VMware did not change. The ESX host does see the increased size however.

1. Is this something i can safely ignore?
2. Have I missed a step?

I reviewed this article below also which doesn't mention anything

VMware Knowledge Base

[tayfundeger]

Did you rescan all after expanding the RDK disk? After expanding the RDM disk on the storage side, you can expand the disk from within the operating system.

[larstr]

KMHH,

Have you tried selecting a host and rescan storage?

Lars

[KMHH]

Thanks for your prompt reply.

Yes I did rescan all HBAs from the ESX host.

It's strange, all seems to be working fine - just the sizes reported by the VM to VMWare (consoled and PowerShell CLI) seems to be wrong and not reporting the extended size of the RDM.

[tayfundeger]

can you connect to esxi with root and check?

If it still looks the same, can you rebuild rdm? Can you review the following KB?

VMware Knowledge Base

[KMHH]

Hi,

I did review that link; had seen it before.  It states the following:

The procedure to expand the size of the RDM depends on the type:

• Physical compatibility mode
  
  Physical compatibility mode RDMs, which are also known as passthru RDMs, expose the physical properties of the mapped LUN to the guest operating system. For the guest operating system to recognize the added space to the expanded mapped LUN, perform a rescan from the ESX host, then from the guest operating system. This process does not require rebooting the virtual machine or the ESX host.
  
  No changes to the RDM files (.vmdk or metadata pointer) are required to take advantage of the added disk space.

My RDM is in Physical compatibility mode so this should apply.

I may put a call in with VMware this afternoon actually to see if I can safely ignore this anomaly.

[KMHH]

Update

It seems that if you run the "Rescan Storage" task twice in quick succession it works.

So, right clicking the ESX host:-  "Storage" - "Rescan Storage" and running twice fixed my issue.

Kath

[tayfundeger]

I suggested that you do rescan first

[KMHH]

I know; but the "single rescan" doesn't work (which I had done before I posted the question); what seems to be key is to run 2 in quick succession - hopefully this should help any other users with the same problem.

Even when VMware were dialled in, they did this and witnessed the same behaviour, they couldn't explain it. 

Just happy that the "double rescan" worked!

Thanks all for posting though, do appreciate the help here.