937948 - Set up holly project branch with nightlies and updates, same builders as m-c

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

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

Per https://msujaws.wordpress.com/2013/11/08/australis-landing-plans/ holly is going to m-c without australis, and nightly updates are going to be needed. Should we also be enabling l10n + l10n nightlies here ? Probably.

Also all the same builders that m-c has, in case we need to use holly to merge to Aurora - we don't want to detect regression at that point.

jaws, what's the timescale here ? See the l10n question above too.

Comment 1

[Jared Wein [:jaws] (please needinfo? me)]

You can see https://mail.mozilla.org/pipermail/firefox-dev/2013-November/001117.html for the latest update on the potential timeline here. 

Thank you for filing this bug. We will need to make sure that the Holly build/l10n environment matches that of mozilla-central.

Comment 2

[Henrik Skupin [:whimboo][⌚️UTC+1]]

We are also planning to include mozmill tests for nightlies on the holly branch. For details see our discussion on https://groups.google.com/forum/#!topic/mozilla.dev.automation/_7mUnkc6468

Comment 4

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

Bug 940382 asks for signed nightly builds for Holly, which is mostly captured here already, but fyi.

Comment 5

[Jared Wein [:jaws] (please needinfo? me)]

Nick, we are now ready to start building signed nightlies. This should exist for the Fx28 Nightly and potentially the Fx29 Nightly.

Comment 6

[Jared Wein [:jaws] (please needinfo? me)]

We don't need l10n builds for holly. Australis added strings but didn't remove any because we expected to back out on Nightly, and the extra CPU cycles needed to run l10n builds doesn't seem worth it.

Comment 7

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

Ok, I'm working on the patch for this.

Comment 8

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

Sorry for the delay here, RelEng is having a work week so we have less time than normal.

Comment 9

[Jared Wein [:jaws] (please needinfo? me)]

Any update on this?

Comment 10

[Aki Sasaki (not active)]

Nick's out sick. Let me see if I can take a stab at it in between crises.

Comment 11

[Aki Sasaki (not active)]

Attachment: holly.diff

Let me know if you want me to move this review elsewhere.  (Or the other one.)

Comment 12

[Aki Sasaki (not active)]

Attachment: holly_aus.diff

go go cvs!
(I did a |php config-dist.php| to make sure I don't blow up AUS like I did with fig.)

Comment 13

[Aki Sasaki (not active)]

Attachment: builderdiff

builder diff

Comment 14

[Aki Sasaki (not active)]

mysql> delete from schedulers where name like '%holly%' and state like '{"last_processed": 2%';
Query OK, 2 rows affected (0.01 sec)

Trying to avoid another scheduler hang...

Comment 15

[Hal Wine [:hwine] use NI!]

Comment on attachment 8338080 [details] [diff] [review]
holly.diff

Review of attachment 8338080 [details] [diff] [review]:
-----------------------------------------------------------------

/me learned about signing server config; lgtm

Comment 16

[Hal Wine [:hwine] use NI!]

Comment on attachment 8338082 [details] [diff] [review]
holly_aus.diff

Review of attachment 8338082 [details] [diff] [review]:
-----------------------------------------------------------------

matches fig

Comment 17

[Aki Sasaki (not active)]

Comment on attachment 8338080 [details] [diff] [review]
holly.diff

https://hg.mozilla.org/build/buildbot-configs/rev/233c56213f44
Pending next merge+reconfig.

Comment 18

[Aki Sasaki (not active)]

Comment on attachment 8338082 [details] [diff] [review]
holly_aus.diff

Landed, but then I thought about Balrog.
Checking in config-dist.php;
/cvsroot/mozilla/webtools/aus/xml/inc/config-dist.php,v  <--  config-dist.php
new revision: 1.279; previous revision: 1.278
done

Comment 19

[Aki Sasaki (not active)]

Ben, Rail, can we get Balrog updates for holly without dealing with the config-dist.php, uh, fun?
I can back out if so.
I can file the blocking IT bug if not.

Comment 20

[[e-mail] (:bhearsum)]

(In reply to Aki Sasaki [:aki] from comment #19)
> Ben, Rail, can we get Balrog updates for holly without dealing with the
> config-dist.php, uh, fun?
> I can back out if so.
> I can file the blocking IT bug if not.

Yup, this is easy, and really the only way to do it. (You'd have to modify in-tree code to point back at aus3 to do it there.) Do you want to learn how to do it, or do you just want someone to do it for you?

Comment 21

[Aki Sasaki (not active)]

(In reply to Ben Hearsum [:bhearsum] from comment #20)
> (In reply to Aki Sasaki [:aki] from comment #19)
> > Ben, Rail, can we get Balrog updates for holly without dealing with the
> > config-dist.php, uh, fun?
> > I can back out if so.
> > I can file the blocking IT bug if not.
> 
> Yup, this is easy, and really the only way to do it. (You'd have to modify
> in-tree code to point back at aus3 to do it there.) Do you want to learn how
> to do it, or do you just want someone to do it for you?

Awesome.  Either works for me.  Learning how probably scales better.

Comment 22

[John Hopkins (:jhopkins)]

in production

Comment 23

[Aki Sasaki (not active)]

https://tbpl.mozilla.org/?tree=Holly&rev=eebfbe8105f5&jobname=nightly
Once those finish, I think I can populate Balrog with the update channel, and resolve this bug.

Comment 24

[Aki Sasaki (not active)]

Added a rule for Firefox-holly-nightly-latest in Balrog.

I think we're done here, although I might back out comment 18.

Comment 25

[Aki Sasaki (not active)]

Backed out:
Checking in config-dist.php;
/cvsroot/mozilla/webtools/aus/xml/inc/config-dist.php,v  <--  config-dist.php
new revision: 1.280; previous revision: 1.279
done