649858 - unify config.py within pools

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

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

I've found a few cases now where slavealloc is pointing a slave at a master that *should* be happy to talk to the master, but it's not.  Mostly this is with staging, but at least one production slave is involved too.  This may be fixed with reconfigs, but will probably also require landing a patch.

Comment 1

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

Attachment: m649858-buildbot-configs-r1.patch

This makes preproduction_config.py match staging_config.py, while adding the missing linux-ix-slave24 to staging_config.py.  Please give this a hard look - I've never quite gotten *_config.py changes right.

Comment 2

[Nick Thomas [:nthomas] (UTC+12)]

Comment on attachment 525890 [details] [diff] [review]
m649858-buildbot-configs-r1.patch

>diff --git a/mozilla/staging_config.py b/mozilla/staging_config.py
>-LINUX_IXS      = ['mv-moz2-linux-ix-slave%02i' % x for x in range(1,24)] + ['linux-ix-slave%02i' % x for x in range(1,43)]
>+LINUX_IXS      = ['mv-moz2-linux-ix-slave%02i' % x for x in range(1,25)] + ['linux-ix-slave%02i' % x for x in range(1,43)]

mv-moz2-linux-ix-slave24 became a win64 ref machine (bug 579331), so we shouldn't add it here or in the preproduction config. r+ if you leave it 'range(1,24)' in both places.

Comment 3

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

Comment on attachment 525890 [details] [diff] [review]
m649858-buildbot-configs-r1.patch

checked in with the requested changes, updated on pp, and reconfig'd (twice).