Greetings,
The short question is: Is there any way to manage Hyper-V between clusters
in a MC configuration without the VMs crashing?
We have seen KBs saying the Witness and persistent file handles can only be
moved between cluster nodes.
We have a MC configuration that was running 9.1P17 and upgraded it to
9.3P14 during a downtime. We also had to physically move one half of the
MC during this downtime.
We have a Hyper-V cluster running Server 2012R2, SMB 3.0, and CIFS file
systems housing the VMs set as continuously available. During the
downtime, most, but not all VM went down hard when switched to the other
half of the MC, and also when switching back. It appears linux VMs housed
on the same cluster/SVM that were moved did not object to the disruption
and did not reboot.
Since the VM disk space went away when cluster1 was switched to cluster2,
the clients had no access to disk space, and we have no event logs. We
assume they paniced and rebooted.
The Hyper-V clients failed immediately as the switchover was started. Are
there any types of timeout settings that can be configured for the VMs so
they just pause during this and pick up again when the SVMs start up on the
other half of the MC?
If the problem is with CIFS, we would be open to iscsi or SAN storage if it
would allow them to survive a MC switchover.
Thanks,
Jeff
--
** Please note my email address has changed to jeff.cleverley(a)broadcom.com
Jeff Cleverley
Factory Systems Engineer
4380 Ziegler Road
Building 1, Dock 1
Fort Collins, Colorado 80525
970-288-4611