1264205 - try all run seems to trigger Android API15+ partner1 opt taskcluster builds

Status: RESOLVED DUPLICATE of bug 1307030

(Firefox Build System :: Task Configuration, task, P4)

Description

[Carsten Book [:Tomcat]]

as example:

https://treeherder.mozilla.org/logviewer.html#?job_id=19402506&repo=try

here it started to trigger a TC  Android API15+ partner1 opt build and i don't think that expected.

Comment 1

[Carsten Book [:Tomcat]]

Greg, Pete do you know if this is expected ?

Comment 2

[Pete Moore [:pmoore][:pete]]

It looks like the partner build was enabled on try in a push on 2015-10-30 17:23 +0000:
https://hg.mozilla.org/try/annotate/508c9dcbeb5ef3490bb1462c29710c83b2db1577/testing/taskcluster/tasks/branches/try/job_flags.yml#l26

I see bug 1163084 landed on Monday - I suspect that was the final piece of getting this in place.

Comment 3

[Pete Moore [:pmoore][:pete]]

Jordan, Nick,

Can you confirm this is expected behaviour?

Many thanks,
Pete

Comment 4

[Greg Arndt [:garndt]]

As far as this repo failing because of a missing cached copy, I have cached it and add it to the list that's automatically cached every 4 hours in bug 1264250

Comment 5

[Jordan Lund (:jlund)]

(In reply to Pete Moore [:pmoore][:pete] from comment #3)
> Jordan, Nick,
> 
> Can you confirm this is expected behaviour?
> 
> Many thanks,
> Pete

iirc, partner builds weren't supposed to be part of `try all`. I suppose one nice side effect is that this decreases the chance of hitting comment 4 as it will refresh the cache :)

I'll let nick make the call if they are ready to have this as regular CI or if they prefer to only run it explicitly.

(In reply to Pete Moore [:pmoore][:pete] from comment #2)

> I see bug 1163084 landed on Monday - I suspect that was the final piece of
> getting this in place.

what part of this bug landed on Monday?

Comment 6

[Nick Alexander :nalexander [he/him]]

(In reply to Jordan Lund (:jlund) from comment #5)
> (In reply to Pete Moore [:pmoore][:pete] from comment #3)
> > Jordan, Nick,
> > 
> > Can you confirm this is expected behaviour?
> > 
> > Many thanks,
> > Pete
> 
> iirc, partner builds weren't supposed to be part of `try all`. I suppose one
> nice side effect is that this decreases the chance of hitting comment 4 as
> it will refresh the cache :)
> 
> I'll let nick make the call if they are ready to have this as regular CI or
> if they prefer to only run it explicitly.

Run it explicitly.

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

we are running this job and it appears to be a perma failed job and in bug 1307030 we determined they are perma failing.

Comment 8

[Dustin J. Mitchell [:dustin] (he/him)]

Yeah, let's nuke the jobs..