2

Environment

Two Nodes A & B

A - Site/Subnet 1

B - Site/Subnet 2

File Share - at Site 3

We have an AG listener that has two IP addresses spanning the two subnets (1 up, the other down depending who is primary)

A - hosts both the primary replica and "Core Cluster Resources"

B - hosts the secondary replica

AG Failover = Automatic Failover

Scenario

Node A is a rebooted

As A is rebooting, I am logged into Node B desktop. In node B, I have the Failover cluster manager open, and the fileshare wintess status is down. This does not make sense, maybe has something to do with Node A owning the "Core Cluster Resources"??????? I would have expected that the file share witness would still show "online"

During this time the AlwaysOn Role in Failover cluster manager is offline, and can not get started since the windows cluster is not online.

Finally after node A completes its reboot cycle, the file share witness is visbile in Failover cluster manager, and the AlwaysOn Role is online. In addition, Node A still owns the "Core Cluster Resources"

I would have expected that the reboot of Node A would have triggered the "Core Cluster Resources" to be owned by Node B (within a few minutes), and the AlwaysOn Role to be brought online by the cluster manager.

I have other alwayson environments (automatic failover, and fileshare witness) that are not multi-subnet, and a reboot of a server hosting a primary replica triggers the cluster manager to bring up the AG on the other node automatically. This behavior is not happening with in a multi subnet environment. Ideas????

My concern is losing the node hosting the primary replica, and the services not being brought up automatically at the second node (Windows and AlwaysON) in this multi-subnet configuration.

Any help would be appreciated.

  • Hi James, have you tried manually failing over to node B and then powering node A off. Just to make sure node B can take ownership and can access the share ok. – Lawrage Feb 11 '16 at 1:19
  • @Lawrage Do you mean moving the core cluster resources over to Node B, rebooting A, while Node A's SQL Instance is hosting the primary replica? – Goforebroke Feb 11 '16 at 4:15

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.