Delayed Failover

[wbratten]

I'm not sure if this is the functionality or I have something configured wrong. I have a primary/active node and secondary/passive node. When I initiate a manual failover or shut down the primary, the secondary takes over as the active node. However, the virtual machine that is running on the ISCSI target will crash during failover. When failover is complete the virtual machine restarts. I would prefer the machine to keep running during failover.

[Robotbeat]

You might need to tighten up the timings on the Open-E side for failover. Also, there may be options on your virtual machine to configure how long to wait for storage to respond. Make sure this is at least 60 seconds. I think it's possible to get the failover time for Open-E to below 30 seconds (or even less), but you should prepare both sides to be as resilient as possible.

[Robotbeat]

Actually, a better time to set up to wait for external storage on your virtual machines should be at least 2 minutes, or 120 seconds.