VM Move fails

Submitted by pcardoso on Mon, 04/07/2014 - 19:36

Hello,

We can't move a specific VM to another Cloudmin server. We get this errors during migration:

Re-fetching system status .. .. done. New status is : Down Refreshing status of host system 77.91.206.10 .. .. done. New status is : Webmin down

Refreshing status of host system pvlc6070.ciberserver.com .. .. done. New status is : Webmin down

.. copy failed : Checksum mismatch after file transfer : 2518d24dceae7e0c6892b58808869fa2 != 8fff3788c7ae7528467dae5d5f0b152e

If we refresh status of the hosts again they recover to webmin (green) If we try to move it again, we get once more.

.. copy failed : Checksum mismatch after file transfer : 35e7925d475d4333cabf911a3848a6f9 != 7d0eb19e0c06ee9adc7031eeccbff7f5

One strange thing is that during this process the VM was offline and even so, the checksum is different.

Both hosts have CentOS 5 in latest version and all updates. They are Xen hosts.

How can we move the VM to the new server?

Status: 
Active

Comments

Submitted by JamieCameron on Mon, 04/07/2014 - 22:25

Are you sure there is no Xen instance accessing the VM disk at the time of the move?

Also, are your VM disks in regular files, on LVM or on an iSCSI server?

Submitted by pcardoso on Tue, 04/08/2014 - 17:22

Hi Jamie,

Nothing access the disks because the move script shuts down the server first.

The VM's are on an LVM volume and I could move two other VM's from the same server to the new server.

Thanks

Submitted by JamieCameron on Tue, 04/08/2014 - 22:29

Is there perhaps another VM that is using the same disk image LV?

Submitted by pcardoso on Wed, 04/09/2014 - 18:17

Hi, No, it's the only VM using the disk. The VM just has that disk and the swap disk but it never migrates the swap because it returns that error first.

Submitted by JamieCameron on Wed, 04/09/2014 - 22:03

Ok, that is very odd. Is there anything different about this VM versus those that move OK?

Also, is the VM shut down when you try the move? You can confirm this with the xm list command on the host system.

Submitted by pcardoso on Thu, 04/10/2014 - 18:55

It's a VM like others and it is shut down. Could it be a transfer error? One thing I see is when the transfer fails, we get a webmin down alert but if we refresh it it detects webmin again.

Submitted by JamieCameron on Thu, 04/10/2014 - 19:11

The fact that Webmin is going down seems suspicious. Does anything get logged to /var/webmin/miniserv.error during the transfer? Also, does the CPU load or RAM usage on the machine spike?