For our SQL DEV/QA environment our system-engineers did setup an Hyper-V enviroment - 3 hosts - 2 VM on each host (1 for DEV an 1 for QA)

Patching for the hosts is scheduled once a month but I noticed the VM's on these hosts are put in a saved mode and not shut down gracefully during the patching.

We are running multiple SQL Server Always On AG's accross these VM. Is it the best practice to shut the VM's down gracefully before rebooting the hosts or can SQL handle the saved state without problems?

Thanks for your help!


Since this not production, it's not so critical to do this the "right" way. Having said that, running QA the same way as production is beneficial for ensuring everything goes well in production.

"Best practices" would be to gracefully shutdown each VM prior to rebooting the host. That allows the AG to realize the databases need to fail over to the remaining node(s). You'd also need to do a single host reboot at a time, so the remaining hosts are able to take the load and retain quorum.

Of course, you could just save-state the VMs and let the AG sort itself out - that is after all what high availability is designed to do.

  • Thanks for this clear answer Max! Just a last question about your last sentence. The save-state shouldn't corrupt the DB's sync state when the VM is resumed? The AG should be able to sort this out itself without intervention? – chittybang Aug 21 at 6:17

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.