2

Problem: In a failover-cluster (Hyper-v) the CSV (iSCSI) gets "pulled" to the node that just restarted (after updates) when i don't want it to.

I can not find any setting defining CSV owner preference, and even if it exists it does not make sense that the volume gets pulled to that node since it was not there when it was rebooted. (It works fine on the VMs)

I have moved everything to the other node (CSV, VMS, Cluster Roles, Witnesses) but when i reboot the "empty" node, it pull the CSV back as soon as it boots.

I have searched and can not find the reason for this.

What can cause the CSV to spontaneously change to that node just after it starts?

Thank you

MaxSantos
  • 21
  • 2
  • Are you pausing the node before you reboot it? If not, you should. That would prevent this from happening in the first place. If you're going to reboot a node or take it out of service for any reason you really should be telling the cluster that by pausing the node. – joeqwerty Jun 06 '18 at 13:48
  • I guess you are right,i should pause the node but that does really not explain why the node that came up "steals" a resource from the other node. It had no resource when it gone down, why does it get one when it comes up. – MaxSantos Jun 06 '18 at 14:03
  • ... And if i have multiple CSV, which ones get move to the new node? – MaxSantos Jun 06 '18 at 14:04
  • Have you found the solution to this problem. It seems that I am in the same boat. The nodes takes the ownership of roles that was just rebooted. Off course, without pausing the roles. But then, other environment doesn't behave like this one. – Sid Mudassir Apr 06 '21 at 18:45

1 Answers1

0

As far as I understand this is "load-balancing" which cannot be disabled in a failover cluster. However you can specify the preferred owner node manually in Failover Cluster Manager "Cluster Properties – General".

BaronSamedi1958
  • 12,510
  • 1
  • 20
  • 46
P.Pauls
  • 266
  • 1
  • 7