HA HyperV - Hosts need updated and restarted, how to manage the guests?

[RabbieBurns]

My HA hyperV cluster hosts need to have windows updates done on them and to be rebooted etc.

 

I have started to live migrate the guests from one to the others so i can then do the updates and restarts etc.

 

However rather than do this all manually is there an easier way to do it? If i just install the updates and then hit reboot will the guests automatically live migrate?

 

Or was my initial thought correct and I need to move them off first?

 

Cheers

[strawberry]

Need to move them first otherwise they just shutdown and reopen on the other node.

[bart21]

should use esx - up on entering maintenance mode on the host it will move all the guests to other hosts in the cluster automatically!

 

 

nick

[ascott2]

System Centre Virtual Machine Manager can put a host into maintenance mode and live migrate to the best possible host.

[RabbieBurns]

If i put it into maintenance mode will it live migrate automatically?

[RabbieBurns]

If i put it into maintenance mode will it live migrate automatically?

 

Yes they do

 

I manually migrated them all first time round, and then after doing updates and restarts on host 1 i put host 2 into maintenance mode and they all migtrated automatically to host 1.

 

However Ive now done updates on host 2 and went to restart but it seems to have hung when restarting at "shutting down cluster service service"

 

Is it safe to cold reboot or should i leave it for a while longer?

[RabbieBurns]

found this page and a fix which has saved a cold reboot

 

Hyper-V Cluster with 2 CSVs hangs on service shutdown of Cluster

 

I just had the same issue today.

 

All virtual machines were stopped before trying to reboot.

 

But instead of cutting the power I tried something a little less drastic.

I tried to kill tasks until the initiated shutdown process could continue.

 

This one worked since I still had ping responses from the server:

taskkill /S vhost2 /IM c* (but you could just try all letters in the alphabet until you get lucky)

 

Result:

SUCCESS: The process "cimlistener.exe" with PID 1592 has been terminated.

SUCCESS: The process "conhost.exe" with PID 1864 has been terminated.

SUCCESS: The process "clussvc.exe" with PID 2148 has been terminated.

 

And then the shutdown completed!

 

Although Im not sure if a subsequent reboot will have the same problem.

[Arthur]

Although I'm not sure if a subsequent reboot will have the same problem.

Do you have any VMs which are configured to use dynamic memory?

 

Hyper-V Host Freezes When Dynamic Memory VM’s Use All RAM

[RabbieBurns]

all of them except the linux guests use dynamic ram.

 

The problem happened exactly as described in the scenario i posted tho rather than any sudden lockup. It was after a maintenance mode reboot after live migrating all the vms from one host to the other.

[RabbieBurns]

this is still happening when I reboot, the cluster service service doesn't stop and i need to use the aforementioned fix to remotely kill the process and then it will shutdown / reboot successfully. Theres a load of posts on technet about this too but none seem to offer a solution...

 

Ive attached the output of cluster.log which is what a lot of the technet articles talk about but not sure if anyone can help...

 

Cluster.zip