Description of problem: In latest snapshot engine again requires bridge 'rhevm' to connect to hosts, this was changed in 3.6 to 'ovirtmgmnt' and should persist between new 3.6.x versions Version-Release number of selected component (if applicable): rhevm-3.6.1.2-0.1.el6.noarch How reproducible: 100% Steps to Reproduce: 1. Have olded snapshot of 3.6 2. Upgrade engine to latest 3.6 Actual results: Host are non-operational as we need to change bridge 'ovirtmgmnt' to bridge 'rhevm' Expected results: Host connection available via bridge 'ovirtmgmnt' Additional info: 2015-12-10 11:11:44,999 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (DefaultQuartzScheduler_Worker-63) [44810b4e] Host 'test' is set to Non-Operational, it is missing the following networks: 'rhevm'
Which 3.6 version did you upgrade from? Could it be that you started with a 3.5 version? When did you add your hosts to your Engine? Which Vdsm version are they running? Did they ever appear as operational? Have you tried it more than once? (if not, the 100% figure is misleading)
I upgraded fresh installation of 3.5 engine. (No hosts in it) Added hosts that were in previous installation of RHEVM3.6 (another older engine). They appeared as non-operational as they did not have correct bridge configured. Tried it on every minor version upgrade as hosts still have the same bridge. This seemes like host installation is creating bridge *rhevm* and reqiures it for host connectivity <- shouldn't that be ovirtmgmnt since 3.6? vdsm-4.17.13-1.el7ev.noarch Resolution here: If we decided to move on 3.6 to use bridge *ovirtmgmnt* this needs fix. If we are keeping the old *rhevm* bridge for 3.6, closed as NOTABUG
We decided to move to ovirtmgmt for fresh 3.6 installations. Upgrades from 3.5 has 'rhevm' as the default. You can (and should) add a new cluster with ovirtmgmt as its management network.