[openstack-dev] [tripleo] Decreasing resource usage in tripleo CI

Derek Higgins derekh at redhat.com
Tue May 6 22:34:41 UTC 2014


Hi,

   I mentioned this at the tripleo meeting today and to get opinions out
there I'd like to bring it up here,

In tripleo we are currently very tight on resources to run our CI jobs
and I'd like to scale back temporarily until we have either more
capacity or reorganize things a bit, to more efficiently use what we
have. Basically I'd like to see us do 3 things

1. Get rid of check-tripleo-seed-precise job
   This is currently a subset of both check-tripleo-undercloud-precise
and check-tripleo-overcloud-precise so is basically using up resources
with no extra gain, if we put plans in place to deploy the
undercloud/overcloud on a persistent seed we can add it back again.

2. Move the check-tripleo-ironic-undercloud-precise into the
experimental-tripleo queue, this test is non voting and currently
doesn't work, when its fixed we can put it back into the check-tripleo
queue and delete the check-tripleo-ironic-seed-precise job, as it should
be a subset of the ironic undercloud job (similar to the seed job
mentioned in point 1)

3. Get back onto 8G nodepool instances, we moved to 16G instances a
while back to work around a problem,
 I've a patch up that should allow us to get back to 8G nodes
https://review.openstack.org/#/c/91545/

Any objections or alternate ideas?

thanks,
Derek.



More information about the OpenStack-dev mailing list