I built a failover cluster on windows 2016 for sql 2016 using SMB fileshare as shared storage. This is just for learning, not production (I'm a SharePoint guy btw)

On the passive node, once installed, the sql service startup parameters were never populated on the server. If I went to the startup parameters in service properties, they were missing. So failing over to this node would never start the server.

Has anyone seen this? I fixed the issue by adding the identical startup params from the active node to the passive node and it all works now.

  • Did you try failing over? This is the way Clusters work. If the parameters on another node were also set to startup, well, you don't want the SQL Service doing that. Since only one node at a time can be active, this is how it should look. – Erik Darling May 2 '17 at 23:16
  • Yeah I tried to fail over but due to naivety, I thought this was damanged but it looks like a lot of what I saw was/is normal. I'm going to build another cluster and test. So question is, on a fresh cluster, prior to the first failover, do the parameters only live on the primary node until failover? Or do both nodes get the parameters? Thanks! – GurdeepS May 2 '17 at 23:36
  • What parameters are you referring to? Specific trace flags? Locations for the error log, and master? – Nic May 3 '17 at 13:01
  • There are settings that are stored in the registry that can get out of sync on the nodes hosting a Failover Clustered Instance (FCI). This doesn't happen often, and tends to be more related to human meddling than anything else, but this blog post outlines a manual way to issue a checkpoint and resync the registry hives between the nodes. – John Eisbrener May 8 '17 at 18:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.