Open Stack

On 2017-01-26 08:45:23 +0100 (+0100), Andreas Jaeger wrote:
[...]
> Yesterday some CI systems needed a restart due to problems caused by one
> provider AFAIK and that lost the post queue. So, this might not have run
> at all.
> 
> If that change is critical, just push a new one up,

Yep, as most jobs that run in post are noncritical (since the next
change which merges will just run them again anyway) we've not made
much effort to find a way to preserve them when we have to perform
emergency restarts of Zuul. That said, if you have one which
actually _was_ critical we can reenqueue the ref for it into the
post pipeline on request.
-- 
Jeremy Stanley