From: "Einav Cohen" <ecohen(a)redhat.com>
To: "users" <users(a)ovirt.org>
Sent: Friday, March 7, 2014 1:24:08 AM
Subject: [Users] [oVirt Test Day 3] oVirt live
based on [1]:
I took the latest (actually, only) ovirt live iso from:
http://jenkins.ovirt.org/view/All/job/ovirt_live_create_iso/
[
http://i.imgur.com/o639iU7.png]
After running the iso (more details to follow), it turns out
that this is a 3.5.0/master iso.
[
http://i.imgur.com/OSO14tu.png]
Google'ing 'ovirt live' didn't bring any 3.4 related results.
I found the GitHub which doesn't seem to have a 3.4 branch
[
http://i.imgur.com/hrsq5ps.png]
and the wiki page [2] that the latest iso in it seems to be
an October 2013 one, which AFAIK is of ovirt-3.3.
details on running the ovirt 3.5.0/master live iso:
I used the 'VM' method.
I used a virt-manager VM on an already well-configured physical
machine, so I skipped the instructions in the 'VM' section [3].
As the ISO seemed to be targeted at el6, I chose in virt-manager
to create a RHEL 6 machine, which is the closest thing to el6
that I could find in virt-manager options.
(no CentOS option in virt-manager, as far as I could see)
[
http://i.imgur.com/fQGGwXK.png]
[ignore the "3.4" in the name of the ISO - that's just the name
that I gave to the ISO file since I wrongfully thought that it
is a 3.4 one]
To allow the VM to function as a hypervisor, I copied the
physical CPU configuration to the VM (which should include
the virtualization-enabled stuff, AFAIK).
[
http://i.imgur.com/JHWwWkp.png]
[I successfully created virtual Hosts a lot of times, so I know
the way to create the VM as I detailed above should work properly]
running the oVirt Live image started off smoothly, problem
encountered when attempting to activate VDSM.
after waiting a lot of time for VDSM to activate, the installation
finally gave up.
[
http://i.imgur.com/tTotnNB.png]
Happened to me too (also while testing oVirt live during the test day).
/var/log/vdsm/vdsm.log was empty and owned by root:root. This makes me
think that the reason is [1].
The image I (and I guess you too) used was built on Mar 5, as you said
it was built from nightly repos, but wrongly used the old nightly repo
(prior to the introduction of /pub/snapshot*). So it included a somewhat
older vdsm - iirc nightly of Feb 26 or so. I was pretty certain that
the vdsm logs issue was solved in master/nightly much before that, but
I might be wrong - [1] points at a lot of changes and I didn't check
all of them. It was moved to ON_QA on Feb 4. I told Ohad about this and
he'll hopefully soon point the ovirt-live build at the new nightly repos.
[1]