Lago jobs migration to jenkins.phx.ovirt.org
David Caro Estevez
dcaro at redhat.com
Wed Apr 6 10:57:15 UTC 2016
On 04/04 23:23, Nadav Goldin wrote:
> updates:
> - lago jobs were migrated, triggering pull request was tested and working,
> lago-bot commenting and check_merged still needs to be tested.
> - 4 new fc23 VMs were added to the new Jenkins instance(fc23-vm04-07)
> - 1 new el7 VM was added (el7-vm25)
>
> I've given admin permissions to all infra members who already enrolled, in
> case anyone needs access.
Things that I had to change so far:
* Allow non-logged in users to read jobs (not managed by puppet)
* Add a new credential for the lago deploy job (ssh to resources.ovirt.org as
lago-deploy-snapshot user, not in puppet, using private key)
* Upgrade the ssh-agent plugin and restart jenkins, as it was not pulling
correctly the upgraded plugin just by 'reloading'
https://gerrit.ovirt.org/#/c/55722/
btw. the jenkins host is currently failing to run puppet (it's using testing
env), so I was unable to actually verify any patches, as I did not want to mess
up any on-going tests
>
>
>
> On Mon, Apr 4, 2016 at 12:16 PM, Nadav Goldin <ngoldin at redhat.com> wrote:
>
> >
> >
> > Just to make sure, don't confuse lago jobs with ovirt-system tests, the
> >> ovirt-system tests use bare metal slaves, those can't be created from
> >> templates
> >> so those can only be installed with pxe (or physically on site, as the
> >> virtual
> >> media of the ilo does not work so good)
> >>
> > sure, only lago* jobs, either way we can have a final check before I add
> > the github credentials.
> >
> > > We already have a template for fc23, it's just creating a new slave
> >> from that
> >> > template (can be done from foreman too, faster than installing from
> >> scratch).
> >
> > thought we could migrate all lago jobs with the slaves at once, but
> > because of [1]
> > NGN needs them too, so either way we need to have fc23 slaves on both
> > jenkins's until
> > full migration.
> >
> > +1, lets try to install a few new f23 slaves from template then.
> >>
> > working on it, will update.
> >
> >
> >
> > [1] https://ovirt-jira.atlassian.net/browse/OVIRT-461
> >
> > On Mon, Apr 4, 2016 at 12:10 PM, Eyal Edri <eedri at redhat.com> wrote:
> >
> >>
> >>
> >> On Mon, Apr 4, 2016 at 11:52 AM, David Caro Estevez <dcaro at redhat.com>
> >> wrote:
> >>
> >>> On 04/04 11:49, Eyal Edri wrote:
> >>> > On Mon, Apr 4, 2016 at 10:38 AM, David Caro Estevez <dcaro at redhat.com>
> >>> > wrote:
> >>> >
> >>> > > On 04/03 20:27, Nadav Goldin wrote:
> >>> > > > Hey David,
> >>> > > > as part of the migration to jenkins.phx.ovirt.org,I want to
> >>> advance with
> >>> > > > the Lago jobs. I already migrated
> >>> > > > infra-puppet/infra-docs/ovirt-node/appliance/imgbased jobs, and so
> >>> far
> >>> > > they
> >>> > > > seem all to work. As far as I understand the Lago jobs are pretty
> >>> > > > independent so it should be rather simple. Currently there are 3
> >>> slaves
> >>> > > > configured (fc23, el7, fc21).
> >>> > >
> >>> > >
> >>> > > There are only 3 fc23 slaves, having one less, duplicates the check
> >>> run
> >>> > > time,
> >>> > > and having only one, triplicates it, can you create new slaves
> >>> instead of
> >>> > > moving them from the old jenkins? (lago is not the only one using
> >>> them, so
> >>> > > migrating all of them is not an option)
> >>> > >
> >>> >
> >>> > Is it possible to add new slaves with the current state of pxe not
> >>> working?
> >>> > The ideal will be to have all new servers installed with the pxe fixed
> >>> so
> >>> > we can deploy many more slaves.
> >>> > This way we can just add lots of slaves to the new jenkins.
> >>>
> >>> We already have a template for fc23, it's just creating a new slave from
> >>> that
> >>> template (can be done from foreman too, faster than installing from
> >>> scratch).
> >>>
> >>
> >> +1, lets try to install a few new f23 slaves from template then.
> >>
> >>
> >>>
> >>> >
> >>> >
> >>> > >
> >>> > >
> >>> > > >
> >>> > > > At the fist stage(until we finish the migration)
> >>> > > > jenkins_master_deploy-configs_merged is not running, so we could
> >>> control
> >>> > > > which jobs get migrated. So if a patch to the jenkins yaml will be
> >>> > > > introduced during the migration process it will have to be re-run
> >>> > > manually.
> >>> > > >
> >>> > > > After migrating I'll disable the lago jobs in jenkins.ovirt.org,
> >>> so
> >>> > > even if
> >>> > > > JJB runs we will have only one jenkins running the CI checks.
> >>> > >
> >>> > > Don't allow both to run anything at the same time, that will lead to
> >>> > > confusion
> >>> > > and branches being deleted at strange times on the github repo, if
> >>> they
> >>> > > run on
> >>> > > one jenkins master, run them there only.
> >>> > >
> >>> > > >
> >>> > > > One more question is if there are any other jobs which are
> >>> dependent on
> >>> > > the
> >>> > > > Lago jobs(like the publishers which are dependent on all
> >>> build_artifacts
> >>> > > on
> >>> > > > ovirt-node/appliance/node)
> >>> > >
> >>> > > Lago is self-contained, anything lago needs (check-build-deploy) is
> >>> tagged
> >>> > > as
> >>> > > lago*, any other job that uses lago, get's it from the repos.
> >>> > >
> >>> > > >
> >>> > > > As far as I understand the only thing needed for migration is
> >>> updating
> >>> > > the
> >>> > > > github api tokens and running JJB with *lago*.
> >>> > >
> >>> > > And disabling the jobs on the other jenkins.
> >>> > > The github configuration is not trivial though, the api token is
> >>> valid only
> >>> > > once and for a specific url. Also you have to configure the github
> >>> hooks to
> >>> > > point to the new jenkins (or it will not get any events), that is
> >>> done at
> >>> > > the
> >>> > > github page, under project configuration.
> >>> > >
> >>> > > >
> >>> > > > What do you think?
> >>> > > >
> >>> > > >
> >>> > > > Thanks
> >>> > > >
> >>> > > > Nadav.
> >>> > >
> >>> > > > _______________________________________________
> >>> > > > Infra mailing list
> >>> > > > Infra at ovirt.org
> >>> > > > http://lists.ovirt.org/mailman/listinfo/infra
> >>> > >
> >>> > >
> >>> > > --
> >>> > > David Caro
> >>> > >
> >>> > > Red Hat S.L.
> >>> > > Continuous Integration Engineer - EMEA ENG Virtualization R&D
> >>> > >
> >>> > > Tel.: +420 532 294 605
> >>> > > Email: dcaro at redhat.com
> >>> > > IRC: dcaro|dcaroest@{freenode|oftc|redhat}
> >>> > > Web: www.redhat.com
> >>> > > RHT Global #: 82-62605
> >>> > >
> >>> > > _______________________________________________
> >>> > > Infra mailing list
> >>> > > Infra at ovirt.org
> >>> > > http://lists.ovirt.org/mailman/listinfo/infra
> >>> > >
> >>> > >
> >>> >
> >>> >
> >>> > --
> >>> > Eyal Edri
> >>> > Associate Manager
> >>> > RHEV DevOps
> >>> > EMEA ENG Virtualization R&D
> >>> > Red Hat Israel
> >>> >
> >>> > phone: +972-9-7692018
> >>> > irc: eedri (on #tlv #rhev-dev #rhev-integ)
> >>>
> >>> > _______________________________________________
> >>> > Infra mailing list
> >>> > Infra at ovirt.org
> >>> > http://lists.ovirt.org/mailman/listinfo/infra
> >>>
> >>>
> >>> --
> >>> David Caro
> >>>
> >>> Red Hat S.L.
> >>> Continuous Integration Engineer - EMEA ENG Virtualization R&D
> >>>
> >>> Tel.: +420 532 294 605
> >>> Email: dcaro at redhat.com
> >>> IRC: dcaro|dcaroest@{freenode|oftc|redhat}
> >>> Web: www.redhat.com
> >>> RHT Global #: 82-62605
> >>>
> >>
> >>
> >>
> >> --
> >> Eyal Edri
> >> Associate Manager
> >> RHEV DevOps
> >> EMEA ENG Virtualization R&D
> >> Red Hat Israel
> >>
> >> phone: +972-9-7692018
> >> irc: eedri (on #tlv #rhev-dev #rhev-integ)
> >>
> >> _______________________________________________
> >> Infra mailing list
> >> Infra at ovirt.org
> >> http://lists.ovirt.org/mailman/listinfo/infra
> >>
> >>
> >
--
David Caro
Red Hat S.L.
Continuous Integration Engineer - EMEA ENG Virtualization R&D
Tel.: +420 532 294 605
Email: dcaro at redhat.com
IRC: dcaro|dcaroest@{freenode|oftc|redhat}
Web: www.redhat.com
RHT Global #: 82-62605
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://lists.ovirt.org/pipermail/infra/attachments/20160406/2da4af43/attachment.sig>
More information about the Infra
mailing list