Cannot migrate vm

[adotjdot]

I just had an issue where I was unable to migrate a vm from one host to the next. I was able to migrate another vm off of the same server just before this one, but this one gets to about 90% then errors out "A general system error occured: Failed waiting for data. Error 16. Invalid argument." The hostd log from that server shows the following during the attempted migration:

Migration changed state from BEGIN to MIGRATING

VMotionCmdCb 'start': done

Disconnect check in progress: /vmfs/volumes/47b0aaa0-6dfb5ada-4527-001b78339936/DIAC-WSUS01U/DIAC-WSUS01U.vmx

Disconnect check in progress: /vmfs/volumes/47b0aaa0-6dfb5ada-4527-001b78339936/DIAC-WSUS01U/DIAC-WSUS01U.vmx

Question info: Error encountered while saving state for migration

, Id: 0 : Type : 2, Default: 0, Number of options: 1

Event generated

Failed to find activation record, event user unknown.

Event 270 : Message on DIAC-WSUS01U on nediacvir0007u.dodiis.mil in ha-datacenter: Error encountered while saving state for migration

FoundryVM_AnswerMessage returned 0

Auto-answered question Error encountered while saving state for migration

Disconnect check in progress: /vmfs/volumes/47b0aaa0-6dfb5ada-4527-001b78339936/DIAC-WSUS01U/DIAC-WSUS01U.vmx

VMotionLastStatusCb: Failed with error 4: Error encountered while saving state for migration.

VMotionResolveCheck: Operation in progress

VMotionStatusCb: Completed

Retrieved current power state from foundry 1

VMotionResolveCheck: Firing ResolveCb

ResolveCb: VMX reports gone = false

ResolveCb: Failed with fault: (vmodl.fault.SystemError) {

reason = "Error encountered while saving state for migration."

msg = ""

}

I can post more if necessary. Any help would be great!

[kjb007]

It may have a pending request. You can try restarting hostd, service mgmt-vmware restart, on that host, and then retry your vmotion.

-KjB

[azn2kew]

Try this solutions it might have in your case since people solved it differently.

http://communities.vmware.com/thread/121727

If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!

Regards,

Stefan Nguyen

iGeek Systems Inc.

VMware, Citrix, Microsoft Consultant

[adotjdot]

I attempted a migration again after restarting the hostd service and the new error was "Ageneral system error occured: Failed to resume on destination." It's the only vm running on that host and it spikes the cpus pretty high when I attempt the migration. I'd like to just shut the vm down and do a cold migration but the server is in production.

[donaldmickey]

Hi,

Look at, the allocation of the affinity of your CPU VM and RAM

[kjb007]

Check your vmware.log again. Also, did you restart hostd on the target host as well?

-KjB

[adotjdot]

No affinity settings. The vm actually ended up powering down so I thought I would try to cold migrate it. I migrated it off, tried to boot it back up and it was giving me swap file errors. I can't delete the swap file because it says "device or resource busy". We've tried everything else since then ranging from unregistering to cold migrating the config files to a different LUN. We are now trying to cold migrate the entire VM and all data to a fresh LUN. Something strange we noticed about the directory one of the vmdk's was in is that there are a couple of additional vmdk's with a <vm-name>-000001.vmdk filename. Is this a snapshot extension? I checked the snapshot manager and there were no active snapshots listed. There is also a second vswp file with a ".20385" extension after the filename.

[kjb007]

Those are snapshots, yes. If they are active, but not showing up in snapshot manager, you can combine them into your vmdk, by creating a new snapshot, and then deleting it. That should clear out any existing snapshots.

The vswp files sometimes will not release, so you can not delete it the vm process is either running or kill'd. If the vm is not running, you should be able to rename it to something else to get your vm up and running.

-KjB

[adotjdot]

I just worked it through with tech support. Turns out a process was locked on the ESX host that I was trying to migrate to. After we killed the process, we could boot the VM back up but still couldn't migrate it. We then built out a new VM and pointed it to the existing vmdks. That worked! We could migrate it without issue at that point. So it seems like there might be an issue with the vmx file. I'm going to try to analyze it a bit further to see if I can find a cause for the failure.

[kjb007]

The new vm would not have the info on the snapshots. I would try to clear those out, as those may be causing issues as well.

Either way, glad things are somewhat working.

-KjB

[adotjdot]

Oh yeah, forgot to mention, the tech helped me merge the snapshots from the command line. So I am running the most recent data....

[kjb007]

Then it may just be a stuck handle in either the esx servers or the vc. I'm not sure if you've rebooted all servers involved, but if you have a fix or a workaround, it's not worth the downtime.

-KjB