From: "Yedidyah Bar David" <didi(a)redhat.com>
To: "Sandro Bonazzola" <sbonazzo(a)redhat.com>
Cc: devel(a)ovirt.org, Users(a)ovirt.org
Sent: Monday, November 10, 2014 3:04:13 PM
Subject: Re: [ovirt-users] [ANN] oVirt snapshots available for testing on EL7
----- Original Message -----
> From: "Sandro Bonazzola" <sbonazzo(a)redhat.com>
> To: announce(a)ovirt.org, devel(a)ovirt.org, Users(a)ovirt.org
> Sent: Monday, November 10, 2014 9:47:56 AM
> Subject: [ovirt-users] [ANN] oVirt snapshots available for testing on EL7
>
> The oVirt team is pleased to announce that the 3.5 and master nightly
> snapshot repositories[1] have been update including builds for
> Red Hat Enterprise Linux 7 (or similar) for testing.
>
> Please note that mirrors may need usually one day before being
> synchronized.
> If you want to be sure to use latest rpms and don't want to wait for the
> mirrors,
> you can edit /etc/yum.repos.d/ovirt-3.5-snapshot.repo or
> /etc/yum.repos.d/ovirt-master-snapshot.repo commenting the mirror line and
> removing the comment on baseurl line.
>
> Thanks to everyone who will join us testing oVirt!
> If you're testing oVirt 3.5 nightly snapshot, please add yourself to the
> test
> page [2]
>
> [1]
http://www.ovirt.org/Install_nightly_snapshot
> [2]
http://www.ovirt.org/Testing/oVirt_3.5.1_Testing
Now tried that, to make it work I had to:
1. Add the following dependencies to the epel repo:
python-crypto,python-cheetah,python-ecdsa,python-markdown
Pushed
http://gerrit.ovirt.org/34993 for that
2. Manually download vdsm-jsonrpc-java from jenkins. David Caro already made
it be published in the yum repos, mirrors will probably take time to sync.
This still was not enough - services (engine, fence-kdump-listener etc) were
not configured correctly.
Tried now with latest 3.5 nightly snapshot and it worked.
Note that yum update to it didn't work - I guess it's because services are
configured only on yum install and not yum update, didn't check this.
Also, I needed to start nfs services prior to running engine-setup, or it
would emit what seems like a non-fatal error in the end when running
'exportfs -ar'. Pushed [3] to fix that.
[3]