Hi, I'm getting a failing jenkins job at http://jenkins.ovirt.org/job/vdsm_create_rpms/label=fedora19/954/console.
From the console it doesn't seem related to the patch change. Can u please advice me what is the problem?
Thanks, Maor
On Sun, Jan 05, 2014 at 04:58:04PM +0200, Maor Lipchuk wrote:
Hi, I'm getting a failing jenkins job at http://jenkins.ovirt.org/job/vdsm_create_rpms/label=fedora19/954/console.
From the console it doesn't seem related to the patch change. Can u please advice me what is the problem?
I do not know what's the exact issue this time, but the main problem is that the same slave is used to run both ovirt-3.3 and "master" jobs. The first require vdsm-python-cpopen, and the second - like your job - conflicts with it. As long as we have two different jobs that attempt to modify the local slave rpm repositories, they must never run in parallel on the same slave. Dan.
i think the new python-cpopen version 1.2.3-7 should resolve this, as we've seen in other jobs. e. ----- Original Message -----
From: "Dan Kenigsberg" <danken@redhat.com> To: "Maor Lipchuk" <mlipchuk@redhat.com>, eedri@redhat.com Cc: infra@ovirt.org, "Ohad Basan" <obasan@redhat.com> Sent: Sunday, January 5, 2014 7:45:32 PM Subject: Re: failing jenkins job
On Sun, Jan 05, 2014 at 04:58:04PM +0200, Maor Lipchuk wrote:
Hi, I'm getting a failing jenkins job at http://jenkins.ovirt.org/job/vdsm_create_rpms/label=fedora19/954/console.
From the console it doesn't seem related to the patch change. Can u please advice me what is the problem?
I do not know what's the exact issue this time, but the main problem is that the same slave is used to run both ovirt-3.3 and "master" jobs. The first require vdsm-python-cpopen, and the second - like your job - conflicts with it.
As long as we have two different jobs that attempt to modify the local slave rpm repositories, they must never run in parallel on the same slave.
Dan.
On Sun, Jan 05, 2014 at 12:55:40PM -0500, Eyal Edri wrote:
i think the new python-cpopen version 1.2.3-7 should resolve this, as we've seen in other jobs.
It may solve one case or another, but as long as it is possible for one job to `yum remove` while another job is running, all bets are off.
participants (3)
-
Dan Kenigsberg -
Eyal Edri -
Maor Lipchuk