Closed Bug 1492695 Opened Last year Closed 5 months ago

Intermittent Jit [taskcluster:error] Task aborted - max run time exceeded


(Core :: JavaScript Engine: JIT, defect, P5)




Tracking Status
firefox68 --- fixed


(Reporter: intermittent-bug-filer, Assigned: gbrown)



(Keywords: intermittent-failure, Whiteboard: [stockwell disable-recommended])


(1 file)

Filed by: btara [at]

02:54:40     INFO -  TEST-PASS | tests\jit-test\jit-test\tests\basic\testMathMinMax.js | Success (code 0, args "--no-baseline --no-ion") [18.1 s]
02:54:40     INFO -  {"action": "test_start", "jitflags": "--no-baseline --no-ion", "pid": 4344, "source": "jittests", "test": "basic\\testMathMinMax.js", "thread": "main", "time": 1537412062.613}
02:54:40     INFO -  {"action": "test_end", "extra": {"jitflags": "--no-baseline --no-ion"}, "jitflags": "--no-baseline --no-ion", "message": "Success", "pid": 4344, "source": "jittests", "status": "PASS", "test": "basic\\testMathMinMax.js", "thread": "main", "time": 1537412080.695}
[taskcluster:error] Aborting task...
[taskcluster 2018-09-20T02:54:48.462Z] SUCCESS: The process with PID 1100 (child process of PID 4344) has been terminated.
[taskcluster 2018-09-20T02:54:48.462Z] ERROR: The process with PID 4004 (child process of PID 4344) could not be terminated.
[taskcluster 2018-09-20T02:54:48.462Z] Reason: There is no running instance of the task.

These lastest classifications are all about jittest-1proc android-hw-p2-8-0 Task aborted - max run time exceeded errors, like these tests that are actually perma now
All are timing out after 60mins and started somewhere between the above range. The tests are queued up for a long time and the backfills/retriggers method makes it hard to investigate.

Recent log:
I was looking through the logcats also, couldn't find anything relevant so this is just a shot in the dark for me, maybe this is relevant:
DatabaseProcessor: processLocalDevices: failed to get the network info.

Geoff, Bob could you please take a look over this as see what's going on here? Thank you.

Flags: needinfo?(gbrown)
Flags: needinfo?(bob)
Flags: needinfo?(kwright)

It looks like the mozharness "install" step is hung and tests are not run at all.

This change might be to blame?

I think this was the only thing allowing "--test-suite=jittest-chunked" with suite configuration for "jittest". (Full disclosure: I asked ahal to make this change -- sorry!)

Let's check...

I'm still waiting on that try push, but should be able to sort this out.

Assignee: nobody → gbrown
Flags: needinfo?(kwright)
Flags: needinfo?(gbrown)
Flags: needinfo?(bob)

Yeah, there's a backlog on those jobs and only 16 devices they run on. Today on autoland I waited around 300 mins just for the jobs to start. Thanks for the try push.

I still haven't managed to verify this on try, but it seems like the best
explanation for the timeouts.

Pushed by
Fix android-hw jittest suite name to avoid timeouts; r=bc
Closed: 5 months ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla68
Regressions: 1546922
Regressions: 1250737
You need to log in before you can comment on or make changes to this bug.