We run a 3-node hyperconverged storage spaces direct cluster with 3-way mirror resiliency. When one node is put into maintenance mode, everthing is still working fine (no errors or inaccessible resources). Shortly after the node is then shutdown, the CSV goes offline with event IDs 5142 and 5150. It seems to me that maintenance mode only affects cluster roles but not the actual physical disks. On the other hand, that shouldn't be a problem in the first place, since 2 more copies are online. Has anyone experienced a similar behaviour or even has a solution/workaround?
-
Microsoft started their hyperconverged game with 4-node as a minimal deployment for a good reason: anything below 4 nodes isn't really reliable. My suggestion would be a) add +1 node to your 3-node S2D HCI cluster, and b) add external SMB3 share as a witness.– BaronSamedi1958Mar 25, 2018 at 15:47
Add a comment
|
1 Answer
It looks like S2D issue specified here: https://social.technet.microsoft.com/Forums/en-US/74dbd313-741e-424e-8acb-d7d68e22174c/two-node-s2d-cluster-disk-volume-fails-when-one-host-goes-down?forum=ws2016 and here:
So, please check the Witness configuration as well as if you have installed the latest Microsoft updates.