libvirt users - Feb 2016 - unhelpful error message on failed "virsh migrate"

Hello,

I have two servers where I can push VMs from one to the other by issuing
the command

 virsh migrate --live --persistent --copy-storage-all --verbose \
        test6 qemu+ssh://kvmhost2/system

on kvmhost1. I can get the VM back by issuing the equivalent command on
kvmhost2:

 virsh migrate --live --persistent --copy-storage-all --verbose \
        test6 qemu+ssh://kvmhost1/system


After an update (on behalf of the glibc bug) it only works for some VMs
and fails for others. The error message is unhelpful:
  
  virsh --debug 4 migrate --live --persistent --copy-storage-all --verbose \
        vmware-mgmt qemu+ssh://kvmhost1/system
error: internal error: info migration reply was missing return status

Is there a way to get a more helpful error message?
migrating VMs only works for a very limited subset of the VMs on the
system in the direction kvmhost2 --> kvmhost2

Stopping the VMs, copying them over to kvmhost1 and restarting still
works, and I can even migrate them back.

How do I debug this problem?

One possible way would be to stop all VMs on kvmhost2, manually copy
them to kvmhost1 and start them
and afterwards reinstall all packages
(or the whole Server if that doesn't help)

But that would be a workaround, and I still would not know what the
problem is.


The system is CentOS 7.2 now and was CentOS 7.1 before the upgrade, with
the qemu-kvm-rhev repo added for qemu.


					Mit freundlichen Gruessen
						Andreas Buschmann
-- 

Andreas Buschmann
[Senior Systems Engineer]

net.DE AG

On Wed 24-Feb-2016 14:34 CET, Andreas Buschmann <buschman at tech.net.de>
wrote:
> Hello,
>
> I have two servers where I can push VMs from one to the other by issuing
> the command
>
>  virsh migrate --live --persistent --copy-storage-all --verbose \
>         test6 qemu+ssh://kvmhost2/system
>
> on kvmhost1. I can get the VM back by issuing the equivalent command on
> kvmhost2:
>
>  virsh migrate --live --persistent --copy-storage-all --verbose \
>         test6 qemu+ssh://kvmhost1/system
Sometimes I get the error message:
error: internal error: info migration reply was missing return status

This usually happens if the target system runs out of space on the
target filesystem.

I can resolve the issue by allocating more space to the target filesystem.


It would be very helpful if there was an error message
error: migrate failed because the target filesystem has insufficient space

This error should go back on stderr, but if that is not possible it
would be sufficient if the error message goes to a logfile or a log
service.

The system is CentOS 7.2, with the qemu-kvm-rhev repo added for qemu.


I also have a pair of test servers (different hardware, same software)
if you want me to run some specific tests.


                                        Mit freundlichen Gruessen
						Andreas Buschmann
-- 

Andreas Buschmann
[Senior Systems Engineer]

net.DE AG