1672162 - TSan needs higher timeout for mochitest-browser

Status: RESOLVED FIXED

(Testing :: Mochitest, enhancement)

Description

[Christian Holler (:decoder)]

In order to successfully run mochitest-browser-chrome with TSan, we need to increase the timeout for these tests, as TSan is even slower than ASan.

Comment 1

[Christian Holler (:decoder)]

Attachment: Bug 1672162 - Increase mochitest-browser timeout for TSan. r?ahal

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/3d04cbe26da4
Increase mochitest-browser timeout for TSan. r=ahal

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/53523cbbc6cc
Fix lint failure in testing/mochitest/runtests. a=lint-fix CLOSED TREE

Comment 4

[Alexandru Michis [:malexandru]]

https://hg.mozilla.org/mozilla-central/rev/3d04cbe26da4
https://hg.mozilla.org/mozilla-central/rev/53523cbbc6cc

Comment 5

[Alexandru Ionescu (needinfo me) [:alexandrui]]

== Change summary for alert #27692 (as of Mon, 16 Nov 2020 15:11:54 GMT) ==

Regressions:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
571%	decision		gecko-decision		220.75 -> 1,481.50

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=27692

Comment 6

[Christian Holler (:decoder)]

(In reply to Alexandru Ionescu (needinfo me) [:alexandrui] from comment #5)

== Change summary for alert #27692 (as of Mon, 16 Nov 2020 15:11:54 GMT) ==

Regressions:

Ratio	Suite	Test	Platform	Options	Absolute values (old vs new)
571%	decision		gecko-decision		220.75 -> 1,481.50

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=27692

This is not possible from the change made.

Comment 7

[Alexandru Ionescu (needinfo me) [:alexandrui]]

Ahal, can you please give an input here? Not really familiar with decision tests, not sure if the backfill applies to them.
Did a rount of backfill, though.

Comment 8

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

The fluctuations in the run time of decision tasks are highly dependent on the network requests made by the scheduler (bugbug) in the decision task. This should be regarded as noise.