Does retrying migrate, one VM at a time, works?
if there is a network throughput problem that would imply deployment problems. Can you measure the capacity of the line between hosts? Also, separating migration network from management network may help to isolate/shape traffic accordingly Is concurrent migration needeD? If there is not enough capacity it owuld be best to configure max_outgoing_migrations to 1 in vdsm.conf on these hypervisors
(In reply to Michal Skrivanek from comment #16) > if there is a network throughput problem that would imply deployment > problems. > Can you measure the capacity of the line between hosts? > Also, separating migration network from management network may help to > isolate/shape traffic accordingly > > Is concurrent migration needeD? If there is not enough capacity it owuld be > best to configure max_outgoing_migrations to 1 in vdsm.conf on these > hypervisors ->Hi, Thank you for your support. Well, our customer uses gigabit network then he thinks that there is no any network traffic problem. And could you inform how to check capacity of line between hosts ? >best to configure max_outgoing_migrations to 1 in vdsm.conf on these > hypervisors -> Please inform its detailed setting for request this thing to the customer. Thank you and regards.
as you say in previous comments, hosts is overloaded when migration starts (even becomes nonresponsive). So the link between those 2 servers is over-utilized and it affects the management communication. Since single VM migration works I would suggest to limit the number of concurrent migrations and/or change bandwidth of each migration Please see https://access.redhat.com/solutions/744423
in the meantime I'm going to move the issue to general migration enhancements planned for 4.0 *** This bug has been marked as a duplicate of bug 1252426 ***