On 26 March 2017 at 19:34, Anantha Raghava
<raghav(a)exzatechconsulting.com> wrote:
Hello,
We are now using oVirt 4.1 in our lab with 2 hosts and a separate server for
engine. Both hosts (Dell R710 Servers) are in a single cluster and load
balancing & migrations work perfectly.
However, if a physical server goes down, the VMs running on that server,
gets paused ans status goes changes to "?". They do not move or restart on
the another host automatically. We have to manually intervene, stop the VMs,
restart them on another server.
Is there a way in which this migration can be automated?
In order for oVirt to migrate VM's it needs to ensure that the VMs are
no longer running on the failed host. Essentially it need to ensure
that host is no longer running (As opposed to not accessible over the
network). You need to configure power management for it to be able to
do that.
(There is a new feature in 4.x to enable storage-based fencing, can't
find a reference for it just now).
--
Barak Korren
bkorren(a)redhat.com
RHCE, RHCi, RHV-DevOps Team
https://ifireball.wordpress.com/