PowerChute 4.1 and DRS

[BillP]

Hi,

Currently running PowerChute 4.1 as a virtual appliance for a 2 host cluster running DRS with VM migration disabled on PCNS and a specific VM shutdown order specified. 

When the "Runtime remaining below threshold" has been triggered, PowerChute starts maintenance mode on each host. When that command is issued to the first host, DRS begins to migrate all the VMs running on that host onto the other host. It then initiates the command to the second host, but this ends up failing after all the VM shutdown timers have elapsed, with the message "The task was canceled by a user" due to it being "Unable to automatically migrate" some VMs from the host.

A few seconds after the maintenance mode commands have been sent, PowerChute attempts to initiate guest OS shutdown for the VMs on the first host. It gives the error message "attempted operation cannot be performed in current state (powered on)" for these VMs because DRS is trying to migrate that VM onto the second host. This results in some VMs not being gracefully shut down as it doesn't make another attempt to shut these VMs after they have successfully migrated due to DRS.

I've read page 29 of the user guide and Application Note 180. It mentions adding a rule to prevent VMs from migrating to the hosts specified in the host group - which is essentially disabling DRS for those VMs. 

Is there any way to keep DRS whilst allowing PowerChute to gracefully shutdown the VMs?

Thanks,

Minh

[BillP]

Hi,

There are 2 options:

You can leave VM Migration disabled in PCNS and change the Automation level in DRS to Partial (it sounds like you have it set to Fully Automated).

Alternatively you can leave DRS set to fully automated and enable VM Migration in PCNS and set a duration to allow enough time for DRS to migrate the VMs before PCNS begins shutting them down. What's happening is PCNS is attempting to shut down a VM that's in the process of migrating and that fails - therefore you need to allow enough time for the VMs to finish migrating (since DRS is fully automated) before the VM shutdown starts.

[BillP]

Thanks Bill. Enabling VM migration and setting the duration ensured that the VMs are shutdown gracefully, however now I am having issues with the high priority VMs powering on.

When PowerChute attempts to issue commands to Power on the High Priority VMs, vCenter shows "A general system error occurred: Connection refused" error. The medium, low and unprioritised VMs do not have any issue powering on. Do you know why is this occurring?

Regards,

Minh

[BillP · Accepted solution]

Hi,

On 10/26/2015 9:42 PM, Minh said:

"A general system error occurred: Connection refused" error

In most cases this indicates a permissions issue. What is displayed in the PowerChute error log? Also, what happens if those VMs are moved to medium or low priority?