Bug 165662 – REGRESSION (r209570-209585): perf-test timing out

• Component: New Bugs
• Status: RESOLVED DUPLICATE

[Ryan Haddad]

Testing runs up until this point:


Running DOM/Accessors.html (37 of 156)

command timed out: 1200 seconds without output running ['python', './Tools/Scripts/run-perf-tests', '--output-json-path', 'perf-test-results.json', '--slave-config-json-path', '../../perf-test-config.json', '--no-show-results', '--reset-results', '--test-results-server', 'perf.webkit.org', '--builder-name', 'Apple Yosemite Release WK2 (Perf)', '--build-number', '6807', '--platform', 'mac-yosemite', '--no-build', '--release'], attempting to kill
process killed by signal 9
program finished with exit code -1
elapsedTime=1913.948860

https://build.webkit.org/builders/Apple%20Yosemite%20Release%20WK2%20%28Perf%29/builds/6807

[Ryan Haddad]

Concurrent GC change is within that range https://trac.webkit.org/changeset/209570

[Filip Pizlo]

I cannot repro just before or just after r209570.

I will keep measuring.  It's possible that my machine can't repro at all.

[Filip Pizlo]

(In reply to comment #2)
> I cannot repro just before or just after r209570.
> 
> I will keep measuring.  It's possible that my machine can't repro at all.

I also could not repro the fail in later revisions on that machine.

It's probably because you need a smaller machine (like a mini) with fewer cores to get the crash.  I'll try on my MBP.

[Alexey Proskuryakov]

Right now, the bot is stuck due to a WindowServer freeze in IOAccelEventMachine2::waitForStamp.

Not sure if it's the root cause though - the current symptom is that testing doesn't start at all, while initially tests were freezing in DOM/Accessors.html.

[Ryan Haddad]

The test is still timing out even with the concurrent GC disabled, so it looks like something else may be at fault:

https://build.webkit.org/builders/Apple%20El%20Capitan%20Release%20WK2%20%28Perf%29/builds/3816

[Filip Pizlo]

(In reply to comment #5)
> The test is still timing out even with the concurrent GC disabled, so it
> looks like something else may be at fault:
> 
> https://build.webkit.org/builders/
> Apple%20El%20Capitan%20Release%20WK2%20%28Perf%29/builds/3816

Could still be my patch, since it was a big one - but it's looking less likely.

[Alexey Proskuryakov]

This test has ceased to time out. Latest timeout was on r209633, then it failed on r209636.

This makes it a duplicate of bug 165693 I think.

*** This bug has been marked as a duplicate of bug 165693 ***