Maintenance mode timeout and retries

[hendersp3]

I am having a situation where maintenance mode is timing out due to a vm not powering down within 300 seconds.  Will ESXi attempt a retry after time out? 

It's funny because I see log entries where it appears to try again, yet other times it does not... Can't seem to get a solid foot on this.  Anyone know for sure?

Thanks

[hussainbte · Accepted solution]

are you referring to VMs migrating when trying to put host into maintenance mode.

you can manually try to migrate them and check if it is failing.

You should have DRS set to fully automated .

To answer the other question , the vCenter will not rerun the maintenance mode task if it fails once.

[hendersp3]

Thank you for the reply.

No I am not referring to vms migrating.  I have a vendor specific appliance that needs to shutdown when the host is entering maintenance mode.  The appliance does shutdown but it takes about 550 seconds therefor the host times out entering into MM. 

The really strange thing is that once the appliance is powered off and I retry to enter MM it still fails with a timeout.  The logs do not reveal anything to me other than it cancelled due to timeout. If I power the appliance on and then back off I can then enter into MM just fine.  Something about the first failure seems to prevent the host from MM for some reason.

Does this sound like anything anyone has seen? 

Thank You

[vijayrana968]

Entering host into maintenance mode may fail due to admission control settings/orphaned VMs on a mounted datastore/if DRS is fully automated or any VM using CD ROM device on host device.

Alternatively, if youre sure if your VM taking 500+ seconds to shutdown, then you can manually enter it using esxcli conveying timeout.

esxcli system maintenanceMode set -e yes -t 600

[hussainbte]

once the timeout value for the maintenance is increase as per the above instructions, I believe you cannot choose to move powered off VMs while entering maintenance.

if it is something that cannot be migrating and taking 500 seconds to shut down, the task cannot wait for more time to complete

[vijayrana968]

You're right. I just tried to isolate this with 2 scenarios.

1. One powered on and one powered off VM on host and I executed the command I mentioned above. Host tried to Enter maintenance mode exact after 10 minutes (600 Sec) but failed with error Time out. Somehow this command is not migrating VMs to another host. Once I triggered it from GUI which asks to migrate VMs in wizard. It completed successfully.