948744 - b2g wasabi build for 1.3

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Aki Sasaki (not active)]

Filing, since bug 909226 got morphed into a one-off 1.2 build bug.

Comment 1

[Jordan Lund (:jlund)]

hmm, looks like I am getting sync errors when running checkout-sources action:
command: script -q -c "/builds/slave/b2g_m-cen_wsb_dep-000000000000/build/repo sync"

log:
10:36:40     INFO - Fetching projects:   5% (5/91)  Fetching project platform/vendor/qcom/msm8960
10:36:41     INFO -  fatal: remote error: FATAL: R any external/caf/platform/vendor/qcom/msm8960 gitweb DENIED by fallthru
10:36:41     INFO -  (or you mis-spelled the reponame)

these seem to be coming from the manifest: https://github.com/mozilla-b2g/b2g-manifest/blob/master/wasabi.xml#L108 <- under wasabi-specific things

Comment 2

[Jordan Lund (:jlund)]

I believe I have wasabi builds running in our buildbot staging environment through my dev machine.

The log here is what I have been able to run:
http://people.mozilla.org/~jlund/b2g_wasabi-110114.txt

There are some errors but I believe those to be upload related steps. Which is to be predicted as my tested slave is using staging keys while trying to push to a production expecting location.

My unknowns:
1) do we want nightlies of this? I have them enabled on my patches but I just wanted to assert.
2) in terms of locales I am guessing we want all locales? As it stands, all locales are enabled
3) Do we want to send uploads to /pvt/mozilla.org/b2gotoro/tinderbox-builds if we are not running a nightly build? As in, should 'wasabi' be in this list -> http://mxr.mozilla.org/build/source/mozharness/configs/b2g/releng-private-updates.py#24

Patches incoming. :)

Comment 3

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-mc-110114.diff

Comment 4

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-bbotcustom-090114.diff

Comment 5

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-bbotcfgs-090114.diff

Comment 6

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-tbpl-100114.diff

Comment 7

[Jordan Lund (:jlund)]

> My unknowns:
> 1) do we want nightlies of this? I have them enabled on my patches but I
> just wanted to assert.
> 2) in terms of locales I am guessing we want all locales? As it stands, all
> locales are enabled
> 3) Do we want to send uploads to /pvt/mozilla.org/b2gotoro/tinderbox-builds

4th unknown:
4) do we need 'engineering' builds of this?

Comment 8

[Ryan VanderMeulen [:RyanVM]]

Comment on attachment 8358871 [details] [diff] [review]
948744_wasabi_1.3_builds-tbpl-100114.diff

Review of attachment 8358871 [details] [diff] [review]:
-----------------------------------------------------------------

r=me with the changes noted

::: js/Config.js
@@ +258,5 @@
>                             "Unagi Device Image Build (Engineering)",
>                             "Unagi Device Image Nightly",
>                             "Unagi Device Image Nightly (Engineering)"],
> +    "Wasabi Device Image": ["Wasabi Device Image Build",
> +                           "Wasabi Device Image Nightly"],

One more space on the second line.

@@ +387,5 @@
>      "Unagi Device Image Build (Engineering)": "Be",
>      "Unagi Device Image Nightly": "N",
>      "Unagi Device Image Nightly (Engineering)": "Ne",
> +    "Wasabi Device Image Build": "B",
> +    "Wasabi Device Image Nightly": "N",

You need a "Wasabi Device Image": "Wasabi" line above these two for the grouping to work.

Comment 9

[Enpei]

Hi,

QA don't need wasabi nightly at this moment since the phone is not for shipping. Chris(catlee) already shared one build for our internal testing. FYI.

Comment 10

[Aki Sasaki (not active)]

Comment on attachment 8358870 [details] [diff] [review]
948744_wasabi_1.3_builds-bbotcfgs-090114.diff

I think the main (only?) difference between releng-private-updates.py and releng-otoro.py is the former allows for updates on nightlies.
Per comment 9, we don't need nightlies, so we probably want to turn them off.

So enable_nightly -> False, and releng-private-updates.py -> releng-otoro.py afaict.

>+# MERGE DAY: wasabi is for B2G 1.3+ only
>+# When gecko29 is on aurora we don't run B2G builds there, but will on beta
>+for _, branch in items_before(BRANCHES, 'gecko_version', 29):
>+        if 'wasabi' in branch['platforms']:
>+                del branch['platforms']['wasabi']

8 spaces -> 4 spaces.

Thanks Jordan!

Comment 11

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-tbpl-130114-2.diff

Thanks for previous review. We no longer require nightlies so I removed some lines and some of my mistakes in last patch are n/a

Comment 12

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-bbotcfgs-130114.diff

using releng-otoro.py, enable_nightly set to False

I will land these patches tonight if r+.

aki, I'm guessing I should land tbpl and m-c patches first? Once m-c patch makes it's way through m-i, I assume it is then that I apply buildbot patches to default? That may be obvious but I just want to assert.

Comment 13

[Ryan VanderMeulen [:RyanVM]]

FYI, you'll need to file a bug to update the production instance of TBPL after you push the TBPL patch (bug 958593 being the most recent instance of such a request). When you push, your patch will show up on http://tbpl-dev.allizom.org/ within 15min or so.

Comment 14

[Aki Sasaki (not active)]

Comment on attachment 8359421 [details] [diff] [review]
948744_wasabi_1.3_builds-bbotcfgs-130114.diff

>+# MERGE DAY: wasabi is for B2G 1.3+ only
>+# When gecko29 is on aurora we don't run B2G builds there, but will on beta
>+for _, branch in items_before(BRANCHES, 'gecko_version', 29):
>+        if 'wasabi' in branch['platforms']:
>+            del branch['platforms']['wasabi']

Four spaces!  The del is good relatively speaking, but the |if 'wasabi'| is 8, so both lines need 4 fewer spaces.

Comment 15

[Jordan Lund (:jlund)]

(In reply to Ryan VanderMeulen [:RyanVM UTC-5] from comment #13)
> FYI, you'll need to file a bug to update the production instance of TBPL
> after you push the TBPL patch

Perfect, thanks! I'll file now.

This has been pushed to default: https://hg.mozilla.org/webtools/tbpl/rev/1ecf06182684

Comment 16

[Ryan VanderMeulen [:RyanVM]]

I'd like to get the TBPL patch for bug 940768 landed before the next push to production as well. If it's OK with you, I'll take care of filing and getting it updated tomorrow.

Comment 17

[Jordan Lund (:jlund)]

mozilla-central patch pushed to inbound: https://hg.mozilla.org/integration/mozilla-inbound/rev/11651cf8335d

Comment 18

[Jordan Lund (:jlund)]

(In reply to Ryan VanderMeulen [:RyanVM UTC-5] from comment #16)
> I'd like to get the TBPL patch for bug 940768 landed before the next push to
> production as well. If it's OK with you, I'll take care of filing and
> getting it updated tomorrow.

No problemo! Whatever works best for you. I won't file a request for my wasabi patch to be merged in production. Leaving it in your hands. Thanks for getting it out in the wild :)

Comment 19

[Ryan VanderMeulen [:RyanVM]]

(In reply to Jordan Lund (:jlund) from comment #15)
> This has been pushed to default:
> https://hg.mozilla.org/webtools/tbpl/rev/1ecf06182684

In production :)

Comment 20

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/11651cf8335d

Comment 21

[Jordan Lund (:jlund)]

buildbot-configs pushed to default: https://hg.mozilla.org/build/buildbot-configs/rev/ca31d149b4f4

buildbotcustom pushed to default: http://hg.mozilla.org/build/buildbotcustom/rev/784510c46bf0

Comment 22

[Aki Sasaki (not active)]

Attachment: fix test-masters.sh

Can't run these on try, anyway.

Comment 23

[Aki Sasaki (not active)]

Comment on attachment 8360039 [details] [diff] [review]
fix test-masters.sh

https://hg.mozilla.org/build/buildbot-configs/rev/60c3a0fe663d

Comment 24

[[e-mail] (:bhearsum)]

something is in production

Comment 25

[Ryan VanderMeulen [:RyanVM]]

Now with less bustage on Aurora. AFAICT, I made all the needed aurora/v1.3 changes to config.json. Would be nice to have a second set of eyes look at it, though.


https://hg.mozilla.org/releases/mozilla-aurora/rev/ae2c436cecdc

Comment 26

[Jordan Lund (:jlund)]

(In reply to Ryan VanderMeulen [:RyanVM UTC-5] from comment #25)
> Now with less bustage on Aurora. AFAICT, I made all the needed aurora/v1.3
> changes to config.json. Would be nice to have a second set of eyes look at
> it, though.

Shoot, I should have realized I'd need to land the patches on aurora too. Thanks for fixing my mistake. /me learns and won't let it happen again.

Comment 27

[Ryan VanderMeulen [:RyanVM]]

FYI, these builds are also running on mozilla-inbound and fx-team, which I don't believe was the intent.

Comment 28

[Jordan Lund (:jlund)]

Attachment: 948744_wasabi_1.3_builds-bbotcfgs-150114.diff

Looks like I enabled this on too many branches. Reverting to the style of explicitness (and 4 spaces!) with only the given branches for now.

Comment 29

[Jordan Lund (:jlund)]

pushed latest buildbot-config patch to default: https://hg.mozilla.org/build/buildbot-configs/rev/1747d962f457

Comment 30

[[e-mail] (:bhearsum)]

something is in production

Comment 31

[Jordan Lund (:jlund)]

These builds are running green on the appropriate branches. Going to mark this as resolved.