[ovirt-users] Linux VMs get stuck after Live Migration

[Soeren Malchow]

Dear all,

some additions:

  *   switching of memory balloonig does not solve the problem
  *   Once we migrated the machines we can basically not use anything anymore that needs some kind of timing – e.g. “ping 1.1.1.1 –i 2” or “while true; do echo “lala”; sleep”; done” both do the frist thing (1 ping and one “lala” output) but then stall

Cheers
Soeren

From: <users-bounces at ovirt.org<mailto:users-bounces at ovirt.org>> on behalf of Soeren Malchow
Date: Wednesday 20 May 2015 13:42
To: "users at ovirt.org<mailto:users at ovirt.org>"
Subject: [ovirt-users] Linux VMs get stuck after Live Migration

Dear all,

We are experiencing a problem with Linux Vms specifically the Vms we tested are CentOS 7.1 (so far it looks as if Windows Vms are not a problem), after the live migration the machines do not show the IP and hostname in the GUI, the same way as if the guest tools are not installed, but the machines are still running.

Then we have 2 different scenarios

Reboot/shutdown:
When i reboot/shutdown one of these VMS from inside the VM the OS seems to shutdown, however, the qemu process is still running and for ovirt the machine is still up, a “shutdown” in the frontend does not do anything, a “power off” shuts down the machine

Just wait:
After a while the machines just seem to shutdown and then basically then same happesn as with reboot/shutdown

The environment is:

Hosted Engine on CentOS 6.6 with ovirt 3.5.2.1
Compute hosts on Fedora 20 with vdsm 4.16.14 and libvirt 1.2.9.1 from the libvirt-preview repo (for live merge)
Storage -> CentOS 7.1 with gluster 3.6.3

Can anyone point us towards the right direction

Regards
Soeren
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20150520/1efd3430/attachment.html>