I've been reading through documentation
https://www.ovirt.org/documentation/architecture/architecture/
https://www.ovirt.org/documentation/self-hosted/Self-Hosted_Engine_Guide/
But am struggling still to understand the role ovirt-engine plays. Would anyone have
recommends for additional reads?
The problem I'm tackling currently looks like this:
- We have (2) oVirt Data Centers, each populated by a single Cluster. The Data Centers
are physically & network-wise 'distant' from one another
- hosted-engine runs on (3) of the (4) Hosts in Data Center A / Cluster A. hosted-engine
does not run on Data Center B / Cluster B
- When we disrupt network connectivity around Cluster A (yes, that's Cluster *A*),
Hosts in Cluster B crash (requiring a power cycle) and Guests in Cluster B get stopped and
paused
I'm struggling to understand why mussing with Cluster A affects Cluster B. From
pcaps, I can see plenty of TLS traffic from Cluster A's Hosts -- presumably from
ovirt-engine running on Cluster A -- exchanged with Cluster B. So, during my last
maintenance window, I put hosted-engine into maintenance mode ... but Hosts/VMs in Cluster
B were still affected.
Where do I go to better understand what ovirt-engine does when it is 'managing'
Hosts & VMs?
--sk