1

This is the error message I get on my replica when there is a connection timeout for few seconds with the Primary. Can anyone tell me whats going on and why are the available storage being moved as this is an AAG where storage is not supposed to be shared and also here there is no failover its just the secondary replica missed heartbeats. Thanks

The failover error logs are mentioned below starting from the bottom

Cluster network 'Cluster Network 2' is operational (up). All available failover cluster nodes attached to the network can communicate using it. Joiner tried to Create Security Context using Package='Kerberos/NTLM' with Context Requirement ='0' and Timeout ='40000' for the target = 'xxxxxx01' The Cluster service successfully brought the clustered role 'Available Storage' offline.

Clustered role 'Available Storage' is moving to cluster node 'xxxx'. Cluster has lost the UDP connection from local endpoint xxx.xx.xx.21:~xx43~ connected to remote endpoint xxx.xx.xx.20:~xx43~. Cluster has missed two consecutive heartbeats for the local endpoint xxx.xx.xx.21:~xx43~ connected to remote endpoint xxx.xx.xx.20:~xx43~.

  • There is an "Available Storage" cluster group, even if there are no resources (shared disks) in it. If there are no resources in it, it doesn't matter if it fails over. Is the availability group being impacted, or you're just wondering why this group is failing over? The real issue, though, is why are you missing heartbeats? You have network issues, buggy NIC drivers, or something going on that needs to be addressed. – Tony Hinkle Oct 27 '17 at 14:21
  • It did not failover. there was a brief outage for few seconds. The missing heartbeats are between Primary replica and secondary replica (asynch mode) that is on another goegraphical location – SqlNovice Oct 27 '17 at 14:34
  • Do you have a file witness set up for a quorum vote? If you have an HQ and a DR site for example, with a 2-node cluster you need a file witness at the HQ site so that the HQ node wins quorum arbitration if you lose network between sites. – Tony Hinkle Oct 27 '17 at 14:53
  • Our environment is like this for the availability group 1) Primary Secondary- SAME Subnet same location and another secondary offsite- and One Listener – SqlNovice Oct 27 '17 at 14:54
  • Since this is not an sql cluster so there is no witness server – SqlNovice Oct 27 '17 at 14:57
1

The only group that failed over was the Available Storage group, which you aren't using, so it is something you can just ignore. If you want to increase the number of missed heartbeats that will cause it to fail over, change the SameSubnetThreshold accordingly.

(get-cluster).SameSubnetThreshold = 15

See Clustering and High-Availability for details.

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.