On Thu, 15 May 2014, Will Dennis wrote:
Hi all,We have been kicking tires on oVirt (was running 3.3.2 up until yesterday; now have upgraded to 3.4.0) and there’s been a parade of bugs… I like the basic functionality, but it just won’t install without a lot of wrangling, or (in some cases) stay running. Now I have a new raft of issues with the 3.4 upgrade as well… this is all running on CentOS 6.5 by the way, on pretty beefy Intel servers. I don’t think I’m up for running this in production, as it may consume a large pie-slice of support time if it continues to give as many problems as it has (granted, it is the “Fedora” of RHEV, so bleeding-edge problems are to be expected I guess.)
Another group at $WORK set up an oVirt cluster (version 3.2.0, on Fedora 18) -- and I'm in the process of assuming management of it.
My experience is that oVirt needs to have its way with a machine; it's been somewhat of a challenge to get it to work and play well with cfengine.
I'm trying to upgrade the nodes to oVirt 3.3 on Fedora 19, and while it works, "works" is very narrowly defined. (Upgrading the Manager will be another project entirely.)
On the other hand, we have two defined clusters, one for centrally managed VMs that require a prescribed amount of CPU/RAM and one in which developers can spin up VMs without regard to contention. The ACLs are pretty cool.
Some things I take for granted in a more traditional KVM environment (like easy access to the VM definitions via virsh) are more difficult since everything is a UUID and stored in a RDBMS. I'm just now learning how to poll Postgres for the info I need for certain reports.
-- Paul Heinlein [email protected] 45°38' N, 122°6' W
_______________________________________________ Tech mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
