So if you’re like me, better safe than sorry… than you’ve probably set your ESX 3.5 U2 HA cluster to “shutdown VM” instead of “Power off VM” or “Leave VM powered on”. By now most of you probably already noticed that when an isolation occurs HA will allow the VM to shutdown clean within 5 minutes. When the 5 minutes are past HA will shutdown the VM’s no matter what.
But for some of you 5 minutes(300 seconds) might just not be long enough, or if you have an ultra fast environment 5 minutes might just be to damn long….
So what can you do to shorten or extent… It’s easy, open up you HA cluster settings and click on advanced options and add the following:
das.isolationShutdownTimeout – values in seconds, default is 300
I’ve also updated my HA advanced settings blog! If anyone has more advanced settings that aren’t on the list let me know!
How does this setting affect the das.failuredetectiontime setting?
We are currently set it to 60000.
So after a minute the ESX server will shutdown all vm’s
if you have your isolation response set to “shutdown guest” than it can take up to 6 minutes before your vm’s are restarted! 1 minute for the detection and a default of 5 for the shutdown of the vm. if after 6 minutes the vm isn’t shutdown by the OS than VMware HA powers it off and restarts it on another host.
Thanks for the update, i had a feeling that it would work that way.
I will test it, to see if i can set it to 0. So that after a minute all vm’s will be shut down, when a ESX server is isolated.
Rob I think you misunderstood my replu. If you set it to 0 than all vm’s will be powered down directly when the isolation response kicks in. if you have your isolation response set to “shutdown vm” instead of “power off vm” you’ll need to have a grace period before it’s being cut off power.