On Wed, Mar 30, 2016 at 10:22 AM, Sharon Naftaly <snaftaly(a)redhat.com> wrote:
On Wed, Mar 30, 2016 at 9:20 AM, Yedidyah Bar David <didi(a)redhat.com> wrote:
> Hi all,
> On Tue, Mar 29, 2016 at 11:12 PM, Jenkins CI <gerrit2(a)gerrit.ovirt.org>
> > Jenkins CI has posted comments on this change.
> > Change subject: packaging: spec: Fix deps for ovirt-engine-tools-backup
> > ......................................................................
> > Patch Set 2:
> > Build Failed
> > : FAILURE
> This failed due to .
> Above run 2500 contained the fix, but still failed. Why? Because it
> started from a broken build. The fix only fixes upgrades from good
> builds to good builds.
> I tried to follow the console logs of various runs of this job and
> have some questions:
> 1. How do we decide what version to upgrade *from*?
The repos we use to upgrade 'from' are:
Afaik the latest version is taken from there (looking at the date modified I
see that yesterday it was indeed
OK, makes sense.
> 2. I suspect that we pick the last "good" version, with some definition of
> but by the time I am looking at this, there are already newer runs
> that succeeded, but all those before 2500 failed. Can't see in the web
> interface older non-failed ones.
> 3. Also searched my mail and could not find a report. When do we send
> email? To whom? Is there an archive somewhere?
In the logs of build #2500 I see
Sending email to: didi(a)redhat.com infra(a)ovirt.org
I meant - do we send also "back to normal" mails? I see some in infra's
archives, but not for this job.
> Not that important, now that all is fixed. Asking mainly for my own
By the way, I see that the 3.6->master upgrade fails on a similar issue:
I know, the patches for 3.6 are pending review.