amurph
Contributor
Contributor

can't power off vm, but the host doesn't think it's running!

Jump to solution

I have two VI3 clusters. An older cluster with 3.02 hosts and a new cluster with ESXi hosts. Current patch levels on the ESXi hosts. The new hosts and old hosts can see the same storage, but I have additional storage that only the new VM's can see.

Most migrations to the new cluster went just fine. A few issues. But one VM is just giving me fits.

After attempts to migrate the vm to new storage failed, I figured I would just migrate the VM to the new hosts and leave it on the old storage for now, until we could figure something out later. Vcenter event logs show that the migration timed out. However, the VM is showing up in the new cluster and is no longer in the old cluster. Also, it powered on all by itself for some reason. Powering off in Vcenter gives the error "Operation failed since another task is in progress." Having had VM's that I could power down in Vcenter before, and not being sure of how to use the RCLI for this, I logged into the host directly in "unsupported mode" and ran vm-support -x to get a list of running vm's and their ID's so that I could kill it. However, the host didn't include this vm in the list of vm's. So I've got a VM that Virtual Center thinks is running, but the host it thinks it's running on doesn't know that so I can't kill the process.

Help?

0 Kudos
1 Solution

Accepted Solutions
kjb007
Immortal
Immortal

Have you tried to restart the management agents from ESXi, as well as restarting the vc service?

-KjB

vExpert/VCP/VCAP vmwise.com / @vmwise -KjB

View solution in original post

0 Kudos
3 Replies
nick_couchman
Immortal
Immortal

Have you rebooted the ESXi server that seems to have the VM in "limbo," yet?

0 Kudos
kjb007
Immortal
Immortal

Have you tried to restart the management agents from ESXi, as well as restarting the vc service?

-KjB

vExpert/VCP/VCAP vmwise.com / @vmwise -KjB

View solution in original post

0 Kudos
amurph
Contributor
Contributor

I must not be thinking today. restarting mgmt agents on host did the trick. Thanks.

0 Kudos