There have been several issues that have been causing sanity jobs for 'master'
review to fail. The CI team believes we now have these resolved.

However, our CI cluster is swamped, and it is likely that we are seeing an 
artificially high failure rate because of the cluster load (some CI sanity
tests are performance-dependent). The CI team is attempting to improve the
situation by managing the cluster load to a more modest level.

Until we get things back to green +1's, we need to take the following actions:

1. Reduce the active reviews that in the queue at any given time (see 
http://www.zuul.opencontrail.org),
ideally to a max of 20 at any given time

2. Reduce the load on CI cluster and jenkins.opencontrail.org by reducing the 
number of 
active build/slave VMs on CI cluster.

3. Priority is for changes required/desired for 3.0. Please
hold off on submitting any reviews that are not required for 3.0.

4. Do not recheck your own reviews. If review is urgent, email 
[email protected] 
and ask us to prioritize it.

We know this is rather drastic, and very inconvenient timing. We appreciate 
your patience
until we get things under control.

--
Klash


_______________________________________________
Dev mailing list
[email protected]
http://lists.opencontrail.org/mailman/listinfo/dev_lists.opencontrail.org

Reply via email to