VM Host Isolation using HA....Query

[Aegis]

Hi All,

I have a quck question related to an ESX cluster HA environment which i could use some clarification on.

I have an environment consisting of 13 ESX servers each in the cluster using HA and DRS. We have suffered a couple of isolation incidents and the guest VM's have not been restarted onto other hosts, as they are meant to.

Looking in the HA configuration the "host isolation response" has been set to "leave VM powered on". During these couple of incidents we noticed that although the hosts were isolated, the guests remained on. However, as the problem with the host was due to a service console memory fault, the host and guests have been permanently isolated and innacesible from the Virtual centre until a full reboot of the ESX server.

My question is, if i change the host isolation response to either shutdown or power off guest, what exactly will happen?? will this either:

1. Simply shut the VM's down and then power them up on a different host?? This is what i thought would happen by default and it is important that they are automatically powered on as we have a lot of guests.

2. Shut the VM's down and leave them on the same host until the host is fixed??

3. Shut the VM's down and move them onto a different host, but then leave them powered off??

I've read lots of documentation but this isn't clear.

Thanks

[weinstein5]

Changing the isolation response to Power Off Guest - will cause the guest to power off on the original ESX Server and which will release the lock on the the virtual disk - the HA cluster will then be able to start the VM on a Node of the cluster -

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful

[Aegis]

OK thats great. So it will definitely power on the guest onto a different host once it has shutdown?

One more thing, our problem i mentioned earlier has caused a few guests to be unable to power on, upon clicking power on the following message appears:

"Could not power onVM: No Swap file. Failed to power on VM."

I've done some research on this and it appears the swap file is locked on another ESX server. We currently have two hosts isolated (as per the previous description) so im expecting its on one of these. Unfortunately i cant restart the hosts until an on-site visit so i could do with another solution if possible.

Do you know if it is possible to either recreate the swap files or delete them??

I have gotten round the problem on one guest by simply creating a new VM and re-using the virtual disk. However, this problem has occured on approx 15-20 guests now so this is prohibitive.

Thanks again,

[riker82]

did you check this thread? http://communities.vmware.com/thread/62945;jsessionid=9B93324D42764EB6BD7FC97AF1CC61BA?tstart=0

[java_cat33]

As Riker82 pointed out in the link - you should check your hosts for any locked processes. You will need to run the following on the hosts

find the process id

ps -axf | grep "vm name"

kill the process

kill -9 processid