VMware Cloud Community
TryllZ
Expert
Expert
‎09-13-2021 10:44 AM

3-Host vSAN Cluster fails to enter Maintenance Mode..

Hi All,

I have a 3 host vSAN cluster, when I want to shutdown all the 3 hosts I run the following code.

Spoiler
Get-Cluster -Name "ClusterC_vSAN" | Get-VMHost | Set-VMHost -State Maintenance -VsanDataMigrationMode NoDataMigration

And what happens is that only 2 hosts enter maintenance mode, while the 3rd one does not and shows the message "Waiting for all VMs to be powered off or suspended or migrated. In a DRS cluster check the Faults page on the DRS tab for troubleshooting." even though al VMs are powered off, and I have specified NoDataMigration, the issue it remains at 17% without progressing.

Progress.png

Yesterday one of the host did not enter Maintenance mode so I had to shut it down gracefully, and now the objects are in a Reduced availability with no rebuild state, I have tried Repair Objects Immediately but the objects have failed and the hosts is no longer part of the vSAN Cluster.

Any thoughts how can I shutdown the hosts safely, and should I remove the host from the Cluster and add it again ?

Thank You

0 Kudos
3 Replies
TheBobkin
Champion
Champion
‎09-13-2021 04:44 PM

@TryllZ , That node is possibly in vSAN Decom state (e.g. Disk-groups and data are not available) but not in vSphere Maintenance Mode - Skyline Health should confirm this - if it is confirmed to be in a vSAN Decom state then migrate any running VMs off it, place it in MM with No Action option then take it out of MM straight after to get MM/Decom state in sync.

0 Kudos
TryllZ
Expert
Expert
‎09-14-2021 01:54 AM

Hi @TheBobkin

Thanks for the reply.

The issue I only have 3 hosts in the vSAN cluster and all 3 need to be shutdown. The 2 can shutdown as they migrate the VM to the 3rd, but the 3rd has no where else to migrate the VM to. And that all VMs are shutdown and no migration set but that does not help. No HA or DRS is configured either.

0 Kudos
depping
Leadership
Leadership
‎09-16-2021 12:58 AM

Check if the vCLS VMs are still up and running, this could stop the process from completing, I have seen that in some cases. if that is the case. Just enable the vCLS Retreat mode, this will delete the VMs. As you are not using HA/DRS, you don't need those VMs for now anyway.

this doc describes how to do it:

https://docs.vmware.com/en/VMware-vSphere/7.0/com.vmware.vsphere.resmgmt.doc/GUID-F98C3C93-875D-4570...

Or watch my video:

https://www.youtube.com/watch?v=WOgkfLL11CY

0 Kudos