1232780 - Schedule mochitest-plain on e10s on Windows 7

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Jonathan Griffin (:jgriffin)]

We should get mochitest-plain running in e10s mode on Windows 7, disabling to get to a green state, while mrbkap continues to work on greening tests.

Comment 1

[Jonathan Griffin (:jgriffin)]

try run to see where we are: https://treeherder.mozilla.org/#/jobs?repo=try&revision=d46bb406d369

Comment 2

[Jonathan Griffin (:jgriffin)]

Oops, that was a bad try syntax for this. Here's a better one: https://treeherder.mozilla.org/#/jobs?repo=try&revision=57953d6975fb

Comment 3

[Jonathan Griffin (:jgriffin)]

new try run with more tests disabled and some assertions marked as expected: https://treeherder.mozilla.org/#/jobs?repo=try&revision=ed0c4866a016

Note that even for passing runs, there are lots of IPC errors in the logs; not sure how important these are.

Comment 4

[Kim Moir [:kmoir] ET]

So these tests be scheduled to run in parallel with the non-e10s tests?  Enabled on a single branch to start?

Comment 5

[Jonathan Griffin (:jgriffin)]

(In reply to Kim Moir [:kmoir] from comment #4)
> So these tests be scheduled to run in parallel with the non-e10s tests? 
> Enabled on a single branch to start?

We're not quite ready to enable them, but when we are, they will be run in parallel with non-e10s and scheduled on all trunk branches, including try.

Comment 6

[Kim Moir [:kmoir] ET]

Vlad or Alin: You could get the patches ready for this bug so the tests could be scheduled as soon as ateam team greens them up

Comment 7

[Vlad Ciobancai [:vladC] [:vciobancai]]

Attachment: bug1232780_config.py.patch

Attached you can find the patch

Comment 8

[Vlad Ciobancai [:vladC] [:vciobancai]]

Attachment: bug1232780_config.py.output

Attached you can find the difference

Comment 9

[Jonathan Griffin (:jgriffin)]

Comment on attachment 8702283 [details] [diff] [review]
bug1232780_config.py.patch

Review of attachment 8702283 [details] [diff] [review]:
-----------------------------------------------------------------

::: mozilla-tests/config.py
@@ +598,4 @@
>      }),
>  ]
>  
> +MOCHITEST_PLAIN = [

We shouldn't define a new suite here; we can use the existing MOCHITEST_E10S suite.

@@ +2404,4 @@
>                      BRANCHES[name]['platforms'][platform][slave_platform]['opt_unittest_suites'] += MOCHITEST_PUSH
>                      BRANCHES[name]['platforms'][platform][slave_platform]['debug_unittest_suites']+= MOCHITEST_PUSH
>  
> +# Bug 1232780 - Schedule mochitest-plain on e10s on Windows 7

This block should be incorporated into the block at http://hg.mozilla.org/build/buildbot-configs/file/tip/mozilla-tests/config.py#l2469, which automatically omits the twig branches.

Comment 10

[Jonathan Griffin (:jgriffin)]

latest try run for test greening, only a few debug failures left: https://treeherder.mozilla.org/#/jobs?repo=try&revision=2a7b3f4af1ca

Comment 11

[Vlad Ciobancai [:vladC] [:vciobancai]]

Attachment: bug1232780_config.py.patch

Updated the patch by using the suggestion received from :jgriffin

Comment 12

[Vlad Ciobancai [:vladC] [:vciobancai]]

Attachment: bug1232780_config.py.output

Attached the difference

Comment 13

[Jonathan Griffin (:jgriffin)]

Comment on attachment 8702483 [details] [diff] [review]
bug1232780_config.py.patch

Review of attachment 8702483 [details] [diff] [review]:
-----------------------------------------------------------------

Looks great, thanks!

Comment 14

[Vlad Ciobancai [:vladC] [:vciobancai]]

(In reply to Jonathan Griffin (:jgriffin) from comment #13)
> Looks great, thanks!

Please tell us when we can push the patches

Comment 15

[Jonathan Griffin (:jgriffin)]

latest try run: https://treeherder.mozilla.org/#/jobs?repo=try&revision=f9230600672c

We're pretty green, although debug chunk 1 seems to suffer from a really high rate of unrelated intermittents. I'm going to retrigger a bunch of times to see if it's green enough to enable.

Comment 16

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/e51048013b92

Comment 17

[Jonathan Griffin (:jgriffin)]

(In reply to Pulsebot from comment #16)
> https://hg.mozilla.org/integration/mozilla-inbound/rev/e51048013b92

This greens up mochitest-plain on opt; I'll land the scheduling changes for this after this is merged to m-c. More work is needed still on debug, which is quite crashy.

Comment 18

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/e51048013b92

Comment 19

[Jonathan Griffin (:jgriffin)]

Comment on attachment 8702483 [details] [diff] [review]
bug1232780_config.py.patch

I landed just the opt part of this as https://hg.mozilla.org/build/buildbot-configs/rev/ea429699060a; the debug part can't land as the tests still need more work there.

Comment 20

[Release Engineering SlaveAPI Service]

In production: https://hg.mozilla.org/build/buildbot-configs/rev/ea429699060a

Comment 21

[Andrew McCreight [:mccr8]]

The two failures in the blocking bugs right now seem to each be confined to specific directories, so we could just disable those for Windows e10s debug runs, to get the rest of the tests going.

Comment 22

[Andrew McCreight [:mccr8]]

(In reply to Pulsebot from comment #16)
> https://hg.mozilla.org/integration/mozilla-inbound/rev/e51048013b92

In the future, it would be good to get some kind of rubber stamp review from somebody responsible for the relevant code when you mass disable tests, so that people working in the area are aware their tests don't work. Thanks!

Comment 23

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/843f2db83213

Comment 24

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/843f2db83213

Comment 25

[Jonathan Griffin (:jgriffin)]

This is long done.

Comment 26

[Firefox Bug Husbandry Bot]

Removing leave-open keyword from resolved bugs, per :sylvestre.