Red Hat Bugzilla – Bug 966587
After upgrading hypervisors from 6.3 to 6.4, spice client gets disconnected upon migration of the VM
Last modified: 2013-11-27 19:05:11 EST
This bug has been copied from bug #920441 and has been proposed
to be backported to 6.4 z-stream (EUS).
We need 2 different networks, the first to be a mgmt one where the
migration will go through, the second where VNC/SPICE clients connect.
However, for better regression testing I suggest testing:
<listen type='address' address='18.104.22.168'/>
<listen type='network' network='net1'/>
<listen type='network' network='net2'/>
where net1 is the one shared with mgm infrastructure, net2 is completely
separated one. Then we should check with default values set in qemu.conf
The example picture of net1 and net2 can be found in the parent bug.
Scratch build created:
Additional patch is needed to fix a crasher bug in the first patch: http://post-office.corp.redhat.com/archives/rhvirt-patches/2013-May/msg00736.html
Created attachment 756169 [details]
Created attachment 756170 [details]
host2 libvirt log
Just for the transcript, as I learned today, "0" *is* a valid IPv4 address. So I need to extend my patch from comment #4 .
Okay, the patch has been proposed on the upstream list:
Moving to POST:
Okay, the previous patch is doomed. Here's a better version:
Created attachment 758656 [details]
host1 libvirtd log
Created attachment 758657 [details]
host 2 libvirtd log
Created attachment 760541 [details]
rhev-h host1 libvirtd log
Created attachment 760542 [details]
rhev-h host2 libvirtd log
Created attachment 760543 [details]
rhel two network host1 libvirtd log
Created attachment 760544 [details]
rhel two network host2 libvirtd log
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.