944100 - b2g 1.2f branch support

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Aki Sasaki (not active)]

Per https://etherpad.mozilla.org/fugu-bring-up .

https://wiki.mozilla.org/ReleaseEngineering:MergeDuty#B2G_Branching has some info, but it's a slightly different process here since it's not a train migration.

We need to branch v1.2f in the b2g repos, a la bug 916911.
Rail has a script: http://hg.mozilla.org/build/braindump/file/a889899ca778/git-related/b2g-branching/b2g_branch_repos.py

The b2g_*.py configs in http://hg.mozilla.org/build/mozharness/file/19ccf160cb4e/configs/b2g need a new version:
Maybe a 'mozilla-b2g26_v1_2f': '1.2.1'  ?
http://hg.mozilla.org/build/mozharness/file/19ccf160cb4e/configs/b2g/releng.py#l45

I think we need a new repo in gaia bumper: http://hg.mozilla.org/build/mozharness/file/321f2b9bc8cd/configs/gaia_bumper/gaia_json.py#l15
we need to poll gaia-1_2f and push to mozilla-b2g26_v1_2f.

Since it looks like we have to provide builds, we do need graphserver to know about this branch, and we do need a tbpl page.  The latter was the long pole for 1.2 branching, so starting earlier might be good.

For l10n I'm leaning towards reusing the same repos as 1.2.

We definitely need all of this in vcs-sync, both legacy and new; our automation depends on both the git and hg sides.

I may be missing something, but this is what's in my head right now.  We can add to the bug or file new as needed.

Comment 1

[Aki Sasaki (not active)]

Landed https://github.com/escapewindow/mozharness/commit/ae29fb725b7a0a5522c6b8c4986f2383516050bb which added a v1.2f branch to https://github.com/escapewindow/test-gecko-git/branches .

Comment 2

[Aki Sasaki (not active)]

Attachment: beagle_1.2f

Basically copies the b2g26_v1_2 block and appends an 'f' to the branch name in 3 places.

Once this lands and merges, it'll be live.  We should get an email to release+vcs2vcs about it within a few minutes.

There's no specific reason to hold off landing.

Comment 3

[Hal Wine [:hwine] use NI!]

Comment on attachment 8339554 [details] [diff] [review]
beagle_1.2f

Review of attachment 8339554 [details] [diff] [review]:
-----------------------------------------------------------------

lgtm

Comment 4

[Aki Sasaki (not active)]

Comment on attachment 8339554 [details] [diff] [review]
beagle_1.2f

https://hg.mozilla.org/build/mozharness/rev/14ea70995c0c
Will go live the next merge.

Comment 5

[Aki Sasaki (not active)]

http://people.mozilla.org/~hwine/tmp/vcs2vcs/add_repo_to_conversion.html for integration/gaia-1_2f using https://hg.mozilla.org/users/hwine_mozilla.com/repo-sync-configs/

Comment 6

[Aki Sasaki (not active)]

Ben: I'm going to guess that setting up the branches in the b2g repos + the b2g-manifests patch should get us past the P1's in https://etherpad.mozilla.org/fugu-bring-up .  vcs-sync may not be immediately required, and builds/tests may be able to wait til next week.  Of course, any progress made on this front would be appreciated...

Comment 7

[[e-mail] (:bhearsum)]

Attachment: add to tbpl

Comment 8

[[e-mail] (:bhearsum)]

Attachment: b2g manifests

I had a few warnings when running b2g_branch_repos.py, I assume they're ignorable but I wanted to run them by you anyways. (Full log available if you want it.):
2013-11-28 09:40:03,980 - WARNING - revision conflict in ['optimus-l5.xml', 'tara.xml']. screencap-gonk: existing rev v1.2, new e6403c71e9eca8cb943739d5a0a192deac60fc51 ['v1.2', 'e6403c71e9eca8cb943739d5a0a192deac60fc51']
2013-11-28 09:40:03,985 - WARNING - revision conflict in ['leo.xml', 'base-jb.xml']. platform_build: existing rev v1.2, new b2g-4.3_r2.1 ['v1.2', 'b2g-4.3_r2.1']
2013-11-28 09:41:18,641 - WARNING - Not updating emulator-jb.xml's platform_external_qemu revision (b2g-jellybean), because it is ponting to a revision different from our branchig point v1.2.
2013-11-28 09:45:29,497 - WARNING - Not updating emulator-jb.xml's android-sdk revision (b2g-4.3_r2.1), because it is ponting to a revision different from our branchig point v1.2.
2013-11-28 09:53:31,198 - WARNING - Not updating base-jb.xml's platform_build revision (b2g-4.3_r2.1), because it is ponting to a revision different from our branchig point v1.2.
2013-11-28 09:53:33,035 - WARNING - Not updating emulator-jb.xml's platform_hardware_ril revision (b2g-jellybean), because it is ponting to a revision different from our branchig point v1.2.
2013-11-28 10:04:17,847 - WARNING - Not updating tara.xml's screencap-gonk revision (e6403c71e9eca8cb943739d5a0a192deac60fc51), because it is ponting to a revision different from our branchig point v1.2.

Comment 9

[[e-mail] (:bhearsum)]

Attachment: mozharness configs

This is just me creating a 1.2f wherever there was a 1.2.

Comment 10

[[e-mail] (:bhearsum)]

Quick status update:

(In reply to Aki Sasaki [:aki] from comment #0)
> We need to branch v1.2f in the b2g repos, a la bug 916911.
> Rail has a script:
> http://hg.mozilla.org/build/braindump/file/a889899ca778/git-related/b2g-
> branching/b2g_branch_repos.py

Repos ready to push, b2g-manifests patch posted.

> The b2g_*.py configs in
> http://hg.mozilla.org/build/mozharness/file/19ccf160cb4e/configs/b2g need a
> new version:
> Maybe a 'mozilla-b2g26_v1_2f': '1.2.1'  ?
> http://hg.mozilla.org/build/mozharness/file/19ccf160cb4e/configs/b2g/releng.
> py#l45
> 
> I think we need a new repo in gaia bumper:
> http://hg.mozilla.org/build/mozharness/file/321f2b9bc8cd/configs/gaia_bumper/
> gaia_json.py#l15
> we need to poll gaia-1_2f and push to mozilla-b2g26_v1_2f.

Patch posted for all mozharness configs.

> Since it looks like we have to provide builds, we do need graphserver to
> know about this branch

> and we do need a tbpl page.  The latter was the long
> pole for 1.2 branching, so starting earlier might be good.

Patch posted.

> For l10n I'm leaning towards reusing the same repos as 1.2.

Etherpad says "Let's not shoot ourselves in the foot on l10n, etc. by making any assumptions about our future support.". This says to me that we want separate l10n branches but I haven't done it yet because I'd rather confirm that first. I assume we can do that later without issue.
 
> We definitely need all of this in vcs-sync, both legacy and new; our
> automation depends on both the git and hg sides.

Haven't touched this yet.

> I may be missing something, but this is what's in my head right now.  We can
> add to the bug or file new as needed.

buildbot-configs still needs to be done, looking out for other things too.

Comment 11

[Ed Morley [:emorley]]

Comment on attachment 8339956 [details] [diff] [review]
add to tbpl

Review of attachment 8339956 [details] [diff] [review]:
-----------------------------------------------------------------

Ty :-)

Comment 12

[[e-mail] (:bhearsum)]

Attachment: hghooks

Comment 13

[[e-mail] (:bhearsum)]

Attachment: production-branches.json

Comment 14

[Rail Aliiev [:rail]]

Comment on attachment 8339957 [details] [diff] [review]
b2g manifests

the patch misses gecko.git and gaia branches, iirc we have to branch them manually...

Comment 15

[Rail Aliiev [:rail]]

Comment on attachment 8339972 [details] [diff] [review]
production-branches.json

this can be landed anytime

Comment 16

[Ed Morley [:emorley]]

Added to treestatus, status set to approval required and MOTD set to same as 1.2
https://treestatus.mozilla.org/mozilla-b2g26_v1_2f

:-)

Comment 17

[[e-mail] (:bhearsum)]

Comment on attachment 8339967 [details] [diff] [review]
hghooks

Need to file a push bug for this still

Comment 18

[[e-mail] (:bhearsum)]

Comment on attachment 8339956 [details] [diff] [review]
add to tbpl

Still need a push bug for this.

Comment 19

[[e-mail] (:bhearsum)]

Comment on attachment 8339958 [details] [diff] [review]
mozharness configs

Landed this on default...will merge to production later.

Comment 20

[[e-mail] (:bhearsum)]

Attachment: update vcs-sync's gecko-git config

Missed this in my earlier patch...

Comment 21

[[e-mail] (:bhearsum)]

Created the Gaia branch:
(b2gbranching)➜  gaia git:(master) git checkout -b v1.2f origin/v1.2
Branch v1.2f set up to track remote branch v1.2 from origin.
Switched to a new branch 'v1.2f'
(b2gbranching)➜  gaia git:(v1.2f) git push --all --dry-run
To [e-mail]:mozilla-b2g/gaia.git
 * [new branch]      v1.2f -> v1.2f
(b2gbranching)➜  gaia git:(v1.2f) git rev-list ^origin/v1.2 v1.2f
(b2gbranching)➜  gaia git:(v1.2f) git rev-list origin/v1.2 ^v1.2f
(b2gbranching)➜  gaia git:(v1.2f) git show-ref v1.2
f13307550640e1ace0f600e6858a73bd6166e801 refs/remotes/origin/v1.2
(b2gbranching)➜  gaia git:(v1.2f) git show-ref v1.2f
f13307550640e1ace0f600e6858a73bd6166e801 refs/heads/v1.2f
(b2gbranching)➜  gaia git:(v1.2f) git push --all
Total 0 (delta 0), reused 0 (delta 0)
To [e-mail]:mozilla-b2g/gaia.git
 * [new branch]      v1.2f -> v1.2f

Rail is working on an update to his b2g branching script because he was concerned about some the warnings I got from the first run.

Comment 22

[[e-mail] (:bhearsum)]

Attachment: b2g manifests

OK, so this is basically s/v1.2/v1.2f/. The only v1.2 left is:
hamachi.xml:  <project path="device/qcom/b2g_common" name="quic/lf/b2g/build" revision="v1.2" />

Which is a repo we can't push to, I think, so we can't change that.

Comment 23

[[e-mail] (:bhearsum)]

OK, I think all branching is done now. Aki did Gecko yesterday, I did Gaia awhile ago, and I just pushed the branches to all the other git repos. Next up: vcs2vcs and vcs-sync.

Comment 24

[[e-mail] (:bhearsum)]

Attachment: vcs2vcs patch #1

Comment 25

[[e-mail] (:bhearsum)]

Attachment: configs for gaia/gecko syncing in vcs2vcs

Comment 26

[[e-mail] (:bhearsum)]

Attachment: graph server stuff

Also adding the plain v1.2 to this config, which never got done in https://bugzilla.mozilla.org/show_bug.cgi?id=932478.

I'll need to run the sql for just the v1.2f stuff by hand.

Comment 27

[[e-mail] (:bhearsum)]

It looks like vcs-sync is working, I see stuff like this in beagle.log:
[[e-mail] vcs2vcs]$ grep 1.2f beagle.log
11:11:01     INFO -                       {'branch_config': {'branches': {'default': 'b2g26_v1_2f'}},
11:11:01     INFO -                        'repo': 'https://hg.mozilla.org/releases/mozilla-b2g26_v1_2f',
11:11:01     INFO -                        'repo_name': 'mozilla-b2g26_v1_2f',
[[e-mail] vcs2vcs]$ grep 2f beagle.log
11:11:01     INFO -                       {'branch_config': {'branches': {'default': 'b2g26_v1_2f'}},
11:11:01     INFO -                        'repo': 'https://hg.mozilla.org/releases/mozilla-b2g26_v1_2f',
11:11:01     INFO -                        'repo_name': 'mozilla-b2g26_v1_2f',
11:11:05     INFO - retry: Calling <bound method HgGitScript.run_command of <__main__.HgGitScript object at 0x2903790>> with args: (['/opt/vcs2vcs/build/venv/bin/hg', '--config', 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1'],), kwargs: {'success_codes': [0, 1, 256], 'error_list': [{'substr': 'command not found', 'level': 'error'}, {'regex': <_sre.SRE_Pattern object at 0x7ff686809570>, 'explanation': 'Automation Error: hg not responding', 'level': 'error'}, {'substr': 'unknown exception encountered', 'explanation': 'Automation Error: python exception in hg', 'level': 'error'}, {'substr': 'failed to import extension', 'explanation': 'Automation Error: hg extension missing', 'level': 'warning'}], 'output_timeout': 300, 'cwd': '/opt/vcs2vcs/build/stage_source/mozilla-b2g26_v1_2f'}, attempt #1
11:11:05     INFO - Running command: ['/opt/vcs2vcs/build/venv/bin/hg', '--config', 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1'] in /opt/vcs2vcs/build/stage_source/mozilla-b2g26_v1_2f
11:11:05     INFO -  comparing with https://hg.mozilla.org/releases/mozilla-b2g26_v1_2f
11:11:05     INFO - No changes for mozilla-b2g26_v1_2f; skipping.

Now looking to make sure vcs2vcs is doing stuff.

Comment 28

[[e-mail] (:bhearsum)]

For the life of me I couldn't make the vcs2vcs docs work (https://people.mozilla.org/~hwine/tmp/vcs2vcs/add_repo_to_conversion.html#setting-up-the-conversion-repo). I got stuck when trying to figure out what to clone. The doc only talks about cloning full repositories - but I clearly don't want every branch of gaia. So instead I rsync'ed integration-gaia-1_2 to integration-gaia-1_2f and updated .git/config and .hg/hgrc to match what was in my patch. I ran do_update by hand and got:
[[e-mail]3 repos]$ do_update git:/opt/vcs2vcs/repos/integration-gaia-1_2f "hg.m.o"
pushing to ssh://hg.m.o/integration/gaia-1_2f
searching for changes
updating bookmark v1.2
exporting bookmark v1.2f
remote: adding changesets
remote: adding manifests
remote: adding file changes
remote: added 3 changesets with 10 changes to 10 files
remote: Trying to insert into pushlog.
remote: Please do not interrupt...
remote: Inserted into the pushlog db successfully.

The "updating bookmark v1.2" worries me slightly, but given that it's only pushing to 1.2f repo, it probably doesn't matter. Hal, can you confirm that?

I still need to add this repo to a job file, patch incoming for that.

Comment 29

[[e-mail] (:bhearsum)]

Attachment: add 1.2f to jobfile

Comment 30

[[e-mail] (:bhearsum)]

Comment on attachment 8340049 [details] [diff] [review]
b2g manifests

I see the v1.2f tags showing up in vcs2vcs logs for these repos now, so I landed this to github.

Comment 31

[[e-mail] (:bhearsum)]

Comment on attachment 8340062 [details] [diff] [review]
graph server stuff

I landed this and ran the 1.2f inserts.

Comment 32

[[e-mail] (:bhearsum)]

Attachment: shorten hamachi

I hit this when running checkconfig with the buildbot-configs I'm about to post.

Comment 33

[[e-mail] (:bhearsum)]

Attachment: build and test configs for 1.2f

This is really  just a copy of all the 1.2 stuff.

Comment 34

[Rail Aliiev [:rail]]

Comment on attachment 8340092 [details] [diff] [review]
build and test configs for 1.2f

looks sane

Comment 35

[Aki Sasaki (not active)]

Comment on attachment 8340092 [details] [diff] [review]
build and test configs for 1.2f

There may be issues with device builds until http://git.mozilla.org/?p=releases/gecko.git;a=summary shows the v1.2f branch... I'm ok waiting for Hal on that one.  Thanks a ton, Ben & Rail!

Comment 36

[Hal Wine [:hwine] use NI!]

(In reply to Ben Hearsum [:bhearsum] from comment #28)
> For the life of me I couldn't make the vcs2vcs docs work
> (https://people.mozilla.org/~hwine/tmp/vcs2vcs/add_repo_to_conversion.
> html#setting-up-the-conversion-repo). I got stuck when trying to figure out
> what to clone. The doc only talks about cloning full repositories - but I
> clearly don't want every branch of gaia. So instead I rsync'ed
> integration-gaia-1_2 to integration-gaia-1_2f and updated .git/config and
> .hg/hgrc to match what was in my patch.

That should work fine. I'll try to clarify the docs -- it's covered as a
"partial conversion" under
https://people.mozilla.org/~hwine/tmp/vcs2vcs/add_repo_to_conversion.html#partial-repos
-- an empty repo is all that is needed.

> I ran do_update by hand and got:
> [[e-mail]3 repos]$ do_update
> git:/opt/vcs2vcs/repos/integration-gaia-1_2f "hg.m.o"
> pushing to ssh://hg.m.o/integration/gaia-1_2f
> searching for changes
> updating bookmark v1.2
> exporting bookmark v1.2f
...
> The "updating bookmark v1.2" worries me slightly, but given that it's only
> pushing to 1.2f repo, it probably doesn't matter. Hal, can you confirm that?

Yes -- because 1_2f is a clone (copy) of 1_2, it has pre-existing v1.2
and master bookmarks (from the git branches of the same name). An empty
repo with a pull from gaia.git would only have the 1.2f, but, as you
note, it's immaterial.

Comment 37

[[e-mail] (:bhearsum)]

(In reply to Aki Sasaki [:aki] from comment #35)
> Comment on attachment 8340092 [details] [diff] [review]
> build and test configs for 1.2f
> 
> There may be issues with device builds until
> http://git.mozilla.org/?p=releases/gecko.git;a=summary shows the v1.2f
> branch... I'm ok waiting for Hal on that one.  Thanks a ton, Ben & Rail!

Hmm, I'm surprised it's not. I'll look into that.

Comment 38

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum [:bhearsum] from comment #37)
> (In reply to Aki Sasaki [:aki] from comment #35)
> > Comment on attachment 8340092 [details] [diff] [review]
> > build and test configs for 1.2f
> > 
> > There may be issues with device builds until
> > http://git.mozilla.org/?p=releases/gecko.git;a=summary shows the v1.2f
> > branch... I'm ok waiting for Hal on that one.  Thanks a ton, Ben & Rail!
> 
> Hmm, I'm surprised it's not. I'll look into that.

It took some sleuthing, but I eventually found this conversation running on github-sync2 under asasaki (!). It looks like it's skipping it because there's nothing new there:
retry: Calling <bound method HgGitScript.run_command of <__main__.HgGitScript object at 0x1936a90>> with args: (['/home/asasaki/gecko/build/venv/bin/hg', '--config', 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1'],), kwargs: {'success_codes': [0, 1, 256], 'error_list': [{'substr': 'command not found', 'level': 'error'}, {'regex': <_sre.SRE_Pattern object at 0x7fdb021c0618>, 'explanation': 'Automation Error: hg not responding', 'level': 'error'}, {'substr': 'unknown exception encountered', 'explanation': 'Automation Error: python exception in hg', 'level': 'error'}, {'substr': 'failed to import extension', 'explanation': 'Automation Error: hg extension missing', 'level': 'warning'}], 'output_timeout': 300, 'cwd': '/home/asasaki/gecko/build/stage_source/mozilla-b2g26_v1_2f'}, attempt #1
Running command: ['/home/asasaki/gecko/build/venv/bin/hg', '--config', 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1'] in /home/asasaki/gecko/build/stage_source/mozilla-b2g26_v1_2f
Copy/paste: /home/asasaki/gecko/build/venv/bin/hg --config web.cacerts=/etc/pki/tls/certs/ca-bundle.crt incoming -n -l 1
Calling ['/home/asasaki/gecko/build/venv/bin/hg', '--config', 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1'] with output_timeout 300
 comparing with https://hg.mozilla.org/releases/mozilla-b2g26_v1_2f
 searching for changes
 no changes found
Return code: 256
No changes for mozilla-b2g26_v1_2f; skipping.

Is that a bug in your script Aki?

I'm going to see if I can force the matter.

Comment 39

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum [:bhearsum] from comment #38)
> (In reply to Ben Hearsum [:bhearsum] from comment #37)
> > (In reply to Aki Sasaki [:aki] from comment #35)
> > > Comment on attachment 8340092 [details] [diff] [review]
> > > build and test configs for 1.2f
> > > 
> > > There may be issues with device builds until
> > > http://git.mozilla.org/?p=releases/gecko.git;a=summary shows the v1.2f
> > > branch... I'm ok waiting for Hal on that one.  Thanks a ton, Ben & Rail!
> > 
> > Hmm, I'm surprised it's not. I'll look into that.
> 
> It took some sleuthing, but I eventually found this conversation running on
> github-sync2 under asasaki (!). It looks like it's skipping it because
> there's nothing new there:
> retry: Calling <bound method HgGitScript.run_command of
> <__main__.HgGitScript object at 0x1936a90>> with args:
> (['/home/asasaki/gecko/build/venv/bin/hg', '--config',
> 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l',
> '1'],), kwargs: {'success_codes': [0, 1, 256], 'error_list': [{'substr':
> 'command not found', 'level': 'error'}, {'regex': <_sre.SRE_Pattern object
> at 0x7fdb021c0618>, 'explanation': 'Automation Error: hg not responding',
> 'level': 'error'}, {'substr': 'unknown exception encountered',
> 'explanation': 'Automation Error: python exception in hg', 'level':
> 'error'}, {'substr': 'failed to import extension', 'explanation':
> 'Automation Error: hg extension missing', 'level': 'warning'}],
> 'output_timeout': 300, 'cwd':
> '/home/asasaki/gecko/build/stage_source/mozilla-b2g26_v1_2f'}, attempt #1
> Running command: ['/home/asasaki/gecko/build/venv/bin/hg', '--config',
> 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1']
> in /home/asasaki/gecko/build/stage_source/mozilla-b2g26_v1_2f
> Copy/paste: /home/asasaki/gecko/build/venv/bin/hg --config
> web.cacerts=/etc/pki/tls/certs/ca-bundle.crt incoming -n -l 1
> Calling ['/home/asasaki/gecko/build/venv/bin/hg', '--config',
> 'web.cacerts=/etc/pki/tls/certs/ca-bundle.crt', 'incoming', '-n', '-l', '1']
> with output_timeout 300
>  comparing with https://hg.mozilla.org/releases/mozilla-b2g26_v1_2f
>  searching for changes
>  no changes found
> Return code: 256
> No changes for mozilla-b2g26_v1_2f; skipping.
> 
> Is that a bug in your script Aki?
> 
> I'm going to see if I can force the matter.

Actually...this looks like it may just be in testing for now, given that remote_targets points at your github repo (https://github.com/mozilla/build-mozharness/blob/master/configs/vcs_sync/gecko-git.py#L188). Back to sleuthing.

Comment 40

[[e-mail] (:bhearsum)]

OK, I think I found it on github-sync3. I pulled ~vcs2vcs/bin, which grabbed attachment #8340054 [details] [diff] [review]. Watching logs now to see if it works...

Comment 41

[[e-mail] (:bhearsum)]

Comment on attachment 8340092 [details] [diff] [review]
build and test configs for 1.2f

Planning to reconfig after I confirm that gecko.git is working correctly.

Comment 42

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum [:bhearsum] from comment #38)
> No changes for mozilla-b2g26_v1_2f; skipping.
> 
> Is that a bug in your script Aki?

I see that your repo has 1.2f in it, so definitely not a bug.

Comment 43

[[e-mail] (:bhearsum)]

We have a 1.2f gecko.git branch \o/: http://git.mozilla.org/?p=releases/gecko.git;a=shortlog;h=refs/heads/v1.2f

Comment 44

[[e-mail] (:bhearsum)]

self serve doesn't have this branch yet...I thought it tracked production-branches.json, but maybe I need to pull that change somewhere before it notices it?

Comment 45

[[e-mail] (:bhearsum)]

Reconfig basically done, kicked a set of dep builds at https://tbpl.mozilla.org/?tree=Mozilla-B2g26-v1.2f&rev=f753e074b175

Comment 46

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum [:bhearsum] from comment #44)
> self serve doesn't have this branch yet...I thought it tracked
> production-branches.json, but maybe I need to pull that change somewhere
> before it notices it?

Self serve is up to date now, but I hit https://bugzilla.mozilla.org/show_bug.cgi?id=944739

Comment 47

[[e-mail] (:bhearsum)]

Getting some graph server post failures during builds:
Executing: ['python', '/builds/slave/m-b26_v1_2f-lx-000000000000000/tools/buildfarm/utils/graph_server_post.py', '--server', 'graphs.mozilla.org', '--selector', '/server/collect.cgi', '--branch', 'Mozilla-B2g26-v1.2f', '--buildid', '20131129080932', '--sourcestamp', 'f753e074b175', '--resultsname', 'Linux_mozilla-b2g26_v1_2f', '--properties-file', 'properties.json', '--timestamp', '1385741110']
results not added, response: 
No machine_name called 'Linux_mozilla-b2g26_v1_2f' can be found
  File "/var/www/html/graphs/server/pyfomatic/collect.py", line 271, in handleRequest
    metadata = MetaDataFromTalos(databaseCursor, databaseModule, inputStream)
  File "/var/www/html/graphs/server/pyfomatic/collect.py", line 63, in __init__
    self.doDatabaseThings(databaseCursor)
  File "/var/www/html/graphs/server/pyfomatic/collect.py", line 92, in doDatabaseThings
    raise DatabaseException("No machine_name called '%s' can be found" % self.machine_name)

Comment 48

[[e-mail] (:bhearsum)]

I think I've got them added now...I don't know why it didn't work the first time though.

Comment 49

[[e-mail] (:bhearsum)]

Yeah, it looks like I didn't add these at all the first time =(. The branch wasn't there and almost caused another failure. Should be fine now though.

Comment 50

[[e-mail] (:bhearsum)]

OK, I'm pretty sure all the builds are working now. https://tbpl.mozilla.org/?tree=Mozilla-B2g26-v1.2f&rev=f753e074b175 has a full set of them. Explanation of the failures there:
* Linux opt purples are because of the graph server issues. They worked after I actually inserted the entries.
* All the blues for Windows Opt are a bad slave that I had to kick.
* The Windows opt orange is a known intermittent failure.
* The Gu red was bustage from https://bugzilla.mozilla.org/show_bug.cgi?id=943008 that I backed out.
* The B2G ICS Emulator Opt orange

I'm pretty sure we're done here now, but I'm leaving this open for someone else (Aki?) to verify on Monday.

Comment 51

[[e-mail] (:bhearsum)]

I happened to be asking about something in #gaia today and Rik told me that he almost deleted the v1.2f branch because he thought it was a typo. He then asked me to send an announcement out to mozilla.dev.gaia, which I did. Is there some sort of process we already have for announcing this sort of thing?

Comment 52

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum [:bhearsum] from comment #50)
> OK, I'm pretty sure all the builds are working now.
> https://tbpl.mozilla.org/?tree=Mozilla-B2g26-v1.2f&rev=f753e074b175 has a
> full set of them. Explanation of the failures there:
> * Linux opt purples are because of the graph server issues. They worked
> after I actually inserted the entries.
> * All the blues for Windows Opt are a bad slave that I had to kick.
> * The Windows opt orange is a known intermittent failure.
> * The Gu red was bustage from
> https://bugzilla.mozilla.org/show_bug.cgi?id=943008 that I backed out.
> * The B2G ICS Emulator Opt orange
> 
> I'm pretty sure we're done here now, but I'm leaving this open for someone
> else (Aki?) to verify on Monday.

No rush Aki, but when you get a chance?

Comment 53

[Aki Sasaki (not active)]

(In reply to Ben Hearsum [:bhearsum] from comment #10)
> > For l10n I'm leaning towards reusing the same repos as 1.2.
> 
> Etherpad says "Let's not shoot ourselves in the foot on l10n, etc. by making
> any assumptions about our future support.". This says to me that we want
> separate l10n branches but I haven't done it yet because I'd rather confirm
> that first. I assume we can do that later without issue.

Yeah, that's why I'm leaning towards no-op right now.

(In reply to Ben Hearsum [:bhearsum] from comment #52)
> (In reply to Ben Hearsum [:bhearsum] from comment #50)
> > I'm pretty sure we're done here now, but I'm leaving this open for someone
> > else (Aki?) to verify on Monday.
> 
> No rush Aki, but when you get a chance?

lgtm.
Thanks a ton!

Comment 54

[Ryan VanderMeulen [:RyanVM]]

Can you please add v1.2 to mcMerge as well?
https://hg.mozilla.org/webtools/tbpl/file/a1502aa96799/mcmerge/js/Config.js#l98

Comment 55

[[e-mail] (:bhearsum)]

Attachment: update mcmerge

I'm assuming we don't want "b2g26" or "mozilla-b2g26" synonyms for v1.2f.

Comment 56

[Ryan VanderMeulen [:RyanVM]]

Comment on attachment 8342355 [details] [diff] [review]
update mcmerge

Definitely not, thanks :)

Comment 57

[[e-mail] (:bhearsum)]

More graph server issues, complaining about a missing Non-PGO: 12:53:40     INFO -  DEBUG : process_Request line: No branch_id for a branch_name 'Mozilla-B2g26-v1.2f-Non-PGO' can be found
12:53:40     INFO -  DEBUG : process_Request line:   File "/var/www/html/graphs/server/pyfomatic/collect.py", line 271, in handleRequest
12:53:40     INFO -  DEBUG : process_Request line:     metadata = MetaDataFromTalos(databaseCursor, databaseModule, inputStream)
12:53:40     INFO -  DEBUG : process_Request line:   File "/var/www/html/graphs/server/pyfomatic/collect.py", line 63, in __init__
12:53:40     INFO -  DEBUG : process_Request line:     self.doDatabaseThings(databaseCursor)
12:53:40     INFO -  DEBUG : process_Request line:   File "/var/www/html/graphs/server/pyfomatic/collect.py", line 113, in doDatabaseThings
12:53:40     INFO -  DEBUG : process_Request line:     raise DatabaseException("No branch_id for a branch_name '%s' can be found" % self.branch_name)

I added the Mozilla-B2g26-v1.2f-Non-PGO branch. Looking to see if there's additional machine entries I'll need.

Comment 58

[[e-mail] (:bhearsum)]

Nothing PGO related in the machines table:
mysql> select * from machines where name like '%PGO%';
Empty set (0.01 sec)

mysql> select * from machines where name like '%pgo%';
Empty set (0.01 sec)



So I think this should be OK now.