1958299 - use t-linux-docker{,-noscratch} pools instead of t-linux-{large,xlarge}{,-source}{,-noscratch}-gcp

Status: RESOLVED FIXED

(Firefox Build System :: Task Configuration, task, P2)

Comment 1

[Julien Cristau [:jcristau] (back on 2026-01-06)]

Attachment: Bug 1958299 - switch most linux tests to new worker pools based on generic-worker + d2g

An exception is made for xpcshell on ccov, due to unresolved timeout /
out of memory issues.

Comment 2

[Julien Cristau [:jcristau] (back on 2026-01-06)]

Attachment: Bug 1958299 - don't retry tasks on SIGKILL.

retry-exit-status 137 was added during the migration from aws to gcp as it was
generally a symptom of spot termination. Since then, worker-runner has learned
to detect when that happens and terminate the task as worker-shutdown, so
this should no longer be necessary when using a recent worker image. It
also interacts badly with how d2g deals with exceeding maxRunTime.

Comment 3

[Julien Cristau [:jcristau] (back on 2026-01-06)]

Attachment: Bug 1958299 - annotate failing tests on new linux CI workers

Comment 4

[Julien Cristau [:jcristau] (back on 2026-01-06)]

I should note this switches some tasks from n2-standard-4 to c2-standard-4 instance types. They should have roughly the same specs though. Previously we had "large" and "xlarge" pools that mapped to basically the same thing (since bug 1862675).

Comment 5

[Julien Cristau [:jcristau] (back on 2026-01-06)]

Attachment: Bug 1958299 - disable webdriver/tests/bidi/network/add_intercept/contexts.py on linux asan. r?whimboo,jdescottes

Comment 6

[Julien Cristau [:jcristau] (back on 2026-01-06)]

Attachment: Bug 1958299 - disable browser_fullscreen_warning.js on linux socketprocess for frequent failures. r?jmaher

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5cbedf2ce275
switch most linux tests to new worker pools based on generic-worker + d2g r=taskgraph-reviewers,bhearsum
https://hg.mozilla.org/integration/autoland/rev/089f4eaad6e6
don't retry tasks on SIGKILL. r=taskgraph-reviewers,Eijebong
https://hg.mozilla.org/integration/autoland/rev/90b2a589155c
disable webdriver/tests/bidi/network/add_intercept/contexts.py on linux asan. r=jdescottes,webdriver-reviewers
https://hg.mozilla.org/integration/autoland/rev/fe6dfa38d472
disable browser_fullscreen_warning.js on linux socketprocess for frequent failures. r=jmaher

Comment 8

[Sandor Molnar[:smolnar]]

https://hg.mozilla.org/mozilla-central/rev/5cbedf2ce275
https://hg.mozilla.org/mozilla-central/rev/089f4eaad6e6
https://hg.mozilla.org/mozilla-central/rev/90b2a589155c
https://hg.mozilla.org/mozilla-central/rev/fe6dfa38d472

Comment 9

[Henrik Skupin [:whimboo][⌚️UTC+1]]

Julien, is D246601 going to be pushed as well? It hasn't been landed yet but this bug is closed as fixed.

Comment 10

[Julien Cristau [:jcristau] (back on 2026-01-06)]

Probably not. I split out the 2 permanent failures, the rest actually seemed to be pre-existing intermittents.

Comment 11

[Acasandrei Beatrice (needinfo me)]

(In reply to Pulsebot from comment #7)

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/5cbedf2ce275
switch most linux tests to new worker pools based on generic-worker + d2g
r=taskgraph-reviewers,bhearsum
https://hg.mozilla.org/integration/autoland/rev/089f4eaad6e6
don't retry tasks on SIGKILL. r=taskgraph-reviewers,Eijebong
https://hg.mozilla.org/integration/autoland/rev/90b2a589155c
disable webdriver/tests/bidi/network/add_intercept/contexts.py on linux
asan. r=jdescottes,webdriver-reviewers
https://hg.mozilla.org/integration/autoland/rev/fe6dfa38d472
disable browser_fullscreen_warning.js on linux socketprocess for frequent
failures. r=jmaher

Perfherder has detected a devtools performance change from push fe6dfa38d472be8cdab3e1c7c74bebe4331d8323.

If you have any questions, please reach out to a performance sheriff. Alternatively, you can find help on Slack by joining #perf-help, and on Matrix you can find help by joining #perftest.

Improvements:

Ratio	Test	Platform	Options	Absolute values (old vs new)
10%	reload-inspector:content-process memory	linux1804-64-qr		8,960,000.00 -> 8,022,357.33
10%	reload-inspector:content-process memory	linux1804-64-qr		9,459,466.24 -> 8,501,248.00

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.

If you need the profiling jobs you can trigger them yourself from treeherder job view or ask a performance sheriff to do that for you.

You can run all of these tests on try with ./mach try perf --alert 44987

The following documentation link provides more information about this command.