Taskcluster failured due to memory "Cannot allocate memory"

[foolip]

The 7 most recent commits to master at the time of writing all have failed Taskcluster runs:
https://tools.taskcluster.net/groups/f9h2rnpNSGSZtf8XqKsI7g
https://tools.taskcluster.net/groups/PEZ79dY9R2GhxheZKTXC3Q
https://tools.taskcluster.net/groups/MEBnx5kARYKxgH42Ot3zsQ
https://tools.taskcluster.net/groups/HiCTerPsTsGZXTaF_Gz4qQ
https://tools.taskcluster.net/groups/Ekt9SdPFT2umvUiNaiTaWQ (for master)
https://tools.taskcluster.net/groups/DWgMz-OFQOKp5iktSe1gSg (for epochs/daily)
https://tools.taskcluster.net/groups/BotCYMVpT--o_Kz1D99puQ
https://tools.taskcluster.net/groups/DevtLTz6TGeBFy8QOFQh8A

All of them have some task that has failed with "Cannot allocate memory" in the log.

Example from the first above:

Traceback (most recent call last):
  File "./wpt", line 5, in <module>
    wpt.main()
  File "/home/test/web-platform-tests/tools/wpt/wpt.py", line 129, in main
    rv = script(*args, **kwargs)
  File "/home/test/web-platform-tests/tools/wpt/run.py", line 540, in run
    rv = run_single(venv, **kwargs) > 0
  File "/home/test/web-platform-tests/tools/wpt/run.py", line 547, in run_single
    return wptrunner.start(**kwargs)
  File "/home/test/web-platform-tests/tools/wptrunner/wptrunner/wptrunner.py", line 331, in start
    return not run_tests(**kwargs)
  File "/home/test/web-platform-tests/tools/wptrunner/wptrunner/wptrunner.py", line 188, in run_tests
    env_extras) as test_environment:
  File "/home/test/web-platform-tests/tools/wptrunner/wptrunner/environment.py", line 61, in __init__
    self.cache_manager = multiprocessing.Manager()
  File "/usr/lib/python2.7/multiprocessing/__init__.py", line 99, in Manager
    m.start()
  File "/usr/lib/python2.7/multiprocessing/managers.py", line 524, in start
    self._process.start()
  File "/usr/lib/python2.7/multiprocessing/process.py", line 130, in start
    self._popen = Popen(self)
  File "/usr/lib/python2.7/multiprocessing/forking.py", line 121, in __init__
    self.pid = os.fork()
OSError: [Errno 12] Cannot allocate memory
Traceback (most recent call last):
  File "./tools/ci/taskcluster-run.py", line 103, in <module>
    main(**vars(parser.parse_args()))
  File "./tools/ci/taskcluster-run.py", line 86, in main
    subprocess.check_call(command)
  File "/usr/lib/python2.7/subprocess.py", line 190, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['python', './wpt', 'run', '--channel=nightly', '--log-wptreport=../artifacts/wpt_report.json', '--no-fail-on-unexpected', '--test-type=testharness', '--this-chunk=14', '--total-chunks=15', '--log-tbpl-level=info', '--log-tbpl=-', '-y', '--no-pause', '--no-restart-on-unexpected', '--install-fonts', '--no-headless', '--install-browser', 'firefox']' returned non-zero exit status 1

[foolip]

It started happening when #14025 was merged. That seems like a somewhat unlikely cause, but I'll try reverting it anyway to see.

[foolip]

Ongoing Taskcluster run with #14025 reverted:
https://tools.taskcluster.net/groups/ZKsyjoJTT3CR9SATJf1RHw

[foolip]

Hmm, maybe it's not so implausible that I caused this. The same log I quoted from also says "No generated manifest found" and generating the manifest does use memory.

[foolip]

https://tools.taskcluster.net/groups/T3FYYkzcSU-j6P8S3qSxFQ is the last good run from before the breakage. There the log says "Manifest downloaded".

[foolip]

Ah... We don't even fetch the tags in the Taskcluster setup:

wpt/tools/docker/start.sh

Lines 27 to 36 in 1ede22e

	# Initially we just fetch 50 commits in order to save several minutes of fetching
	git fetch --quiet --depth=50 origin ${REF}
	if [[ ! `git rev-parse --verify -q ${REVISION}` ]];
	then
	# But if for some reason the commit under test isn't in that range, we give in and
	# fetch everything
	git fetch -q --unshallow ${REMOTE}
	git rev-parse --verify ${REVISION}
	fi

[foolip]

All right, it was my fault, Taskcluster is now passing again.