I need to put a host into maintenance mode in my ESXi 6.7 U3 cluster. All the VMs migrated off of it just fine...except one, which consistently hangs at 21% and then fails.
The error stack in the task says:
And the reason in the relevant event is:
However, this is red herring because, as I said, all the other VMs were able to vMotion without issue.
Looking at the hardware, the only thing unusual I see is that it uses IDE disks, not virtual SCSI, but that shouldn't prevent vMotion, should it?
Any ideas how I can troubleshoot this? Unfortunately, I'm finding nothing useful online from that error stack info because all the advice pertains to when vMotion as a whole doesn't work on a host, which isn't relevant here.
It's not necessarily a red herring. A vMotion operation moves the contents of all memory pages to another eligible host, and if that VM is "dirtying" memory pages at a rate faster than they can be drained, the operation may fail. Is this VM fairly memory intensive?
Is this VM fairly memory intensive?
Hardly at all:
[root@ ~]# free -g
total used free shared buff/cache available
Mem: 15 1 12 0 1 13
Swap: 7 0 7
There are 8 processes using around ~1% of memory each. Everything else is using less than that.
That looks fairly innocent. You may want to inspect the VM's log file to see what gets written after the vMotion fails. Also check the vmkernel.log file on the source host.