Right, and even if we did change all the 3rd parties over to something
that doesn't start with 'recheck' there will be inconsistencies between
3rd parties and 1st party.
If you can retrigger all systems (1st or 3rd) in the same way it'll be
much easier for users. For example, if you see "recheck jenkins"
rerunning the jenkins tests then it follows logicially that "recheck
turbo-hipster" makes that bot re-trigger.
Cheers,
Josh
Rackspace Australia
On 8/12/14 5:27 PM, Ruslan Kamaldinov wrote:
We've reconfigured our ci to be re-triggered only on command
"retrigger murano-ci" for exactly this reason. But, I guess that it'll
take a lot of time to convince all 3rd party CI systems to do the same
and to use the same trigger keyword. So, maybe it'll be easier to
update OpenStack CI configuration the way Joshua suggested.
Thanks,
Ruslan
On Tue, Aug 12, 2014 at 8:15 AM, Joshua Hesketh
<[email protected]> wrote:
Hi Sergey,
Right, what I am actually proposing is to /remove/ that functionality.
Specifically how it is currently configured there is no way to trigger 3rd
party tests without rerunning (and hence wasting) 1st party CI.
Cheers,
Josh
Rackspace Australia
On 8/11/14 9:54 PM, Sergey Lukjanov wrote:
FWIW it's already configured to recheck on recheck .*
https://github.com/openstack-infra/config/blob/master/modules/openstack_project/files/zuul/layout.yaml#L18
On Fri, Aug 8, 2014 at 5:56 AM, Joshua Hesketh
<[email protected]> wrote:
Howdy,
So I'm wondering if we can optimise our resource usage by limiting when
we
run rechecks a little tighter. A recent change[0] removed the need for
supplying 'no bug' or 'bug #' as it was decided these were unnecessary.
However we are now triggering rechecks on 'recheck*' which means when
somebody wants to recheck a third party they will also be wasting
OpenStack's resources (eg 'recheck migrations' will trigger both
turbo-hipster and jenkins).
There is also the case where we might want to recheck Jenkins and not a
third party. Most 1st and 3rd party systems both currently trigger on
'recheck no bug' for legacy reasons but it is wasting both 3rd and 1st
party
resources. Basically we need a way of rechecking only the failing system
or
the system the commenter is concerned with.
I'd like to propose that OpenStack's CI triggers on the following
comments:
recheck no bug
recheck bug #
recheck openstack-ci
recheck jenkins
recheck all
The first two would be there for legacy matching while we re-train users
to
use the latter. The thought process here is that once we figure out the
naming of CI systems we can just use 'recheck system-name' and save on
resources by not rechecking everything all the time. Rechecking based on
the
username commenting on the system is a usability thing in my mind. This
should make it obvious to developers how to target their rechecks.
Thoughts?
Cheers,
Josh
[0] https://review.openstack.org/#/c/108724/
--
Rackspace Australia
_______________________________________________
OpenStack-Infra mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
_______________________________________________
OpenStack-Infra mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
_______________________________________________
OpenStack-Infra mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
