Comment 5 for bug 1826051

Hi Corey,
that is bad news - I literally migrated 10s of thousands of VMs T->B now and newer hit that so far.

A few questions to get an initial hold on this:
1. is libvirt or the actual guest (qemu) going to 100% CPU?
2. can you provide an libvirt XML of one of the failing guests (to check if they have anything special)
3. are one or all thready going insane on 100% CPU, maybe check with
   $ pidstat -t -p <pid-that-hogs> 5 10
4. If it is at 100% it must spin "somewhere" could you take a look with a few tools:
   strace -f -r -T -p <pid-that-hangs>
   If you can install `perf` that would be great as well.