678381 - Intermittent OSError: [Errno 2] No such file or directory: 'browser_output.txt' in Android talos

Status: RESOLVED FIXED

(Release Engineering :: General, defect, P3)

Description

[Daniel Holbert [:dholbert]]

http://tinderbox.mozilla.org/showlog.cgi?log=Mozilla-Inbound/1313095650.1313097145.28811.gz
Android Tegra 250 mozilla-inbound talos remote-ts on 2011/08/11 13:47:30
{
Running test ts: 
		Started Thu, 11 Aug 2011 13:50:37
reconnecting socket
pushing directory: /tmp/tmpx_JiQ7/profile to /mnt/sdcard/tests/profile
Failed ts: 
		Stopped Thu, 11 Aug 2011 14:11:20
FAIL: Busted: ts
FAIL: initialization timed out
Completed test ts: 
		Stopped Thu, 11 Aug 2011 14:11:20
RETURN: cycle time: 00:20:42<br>
tegra-070: 
		Stopped Thu, 11 Aug 2011 14:11:20
Sending results: 
		Started Thu, 11 Aug 2011 14:11:20
RETURN:<br>
RETURN:<p style="font-size:smaller;">Details:<br>|</p>
Completed sending results: 
		Stopped Thu, 11 Aug 2011 14:11:20
reconnecting socket
FIRE PROC: ' "MOZ_CRASHREPORTER_SHUTDOWN=1,MOZ_CRASHREPORTER_NO_REPORT=1,NO_EM_RESTART=1" org.mozilla.fennec  -profile /mnt/sdcard/tests/profile http://bm-remote.build.mozilla.org/getInfo.html'
DeviceManager: error pulling file: No such file or directory
Traceback (most recent call last):
  File "bcontroller.py", line 244, in <module>
    sys.exit(main())
  File "bcontroller.py", line 238, in main
    bcontroller.run()
  File "bcontroller.py", line 163, in run
    os.chmod(self.log, 0777)
OSError: [Errno 2] No such file or directory: 'browser_output.txt'
program finished with exit code 0
elapsedTime=1244.847692
TinderboxPrint:s: tegra-070
}

Looks like bug 665967 tracked a related issue a few months back, but that was RESOLVED|FIXED, so I'm filing this as a new bug.

Comment 1

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

I dont have clear impression of how often this is happening. Should this be marked "[android_tier_1]"?

Comment 2

[Phil Ringnalda (:philor)]

Nobody knows, because nobody knows what "this" is. This symptom of whatever the actual problem is appears frequently, but that may mean it's a symptom of one frequent problem, of ten common problems, or of thirty infrequent problems, or that it's just one of the many symptoms of the one crash that gets the ateam's blame for virtually every Android test failure.

Comment 3

[John Ford [:jhford] CET/CEST Berlin Time]

Do we know how often this symptom is occurring at least?

Comment 4

[Phil Ringnalda (:philor)]

Bwahaha, you just asked for log URL copy-pasting. You'll be soooooorrry. :)

Comment 5

[Phil Ringnalda (:philor)]

Doesn't seem all _that_ frequent, though http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1313777309.1313779259.21941.gz and http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1313777309.1313778838.20529.gz stayed on my screen long enough for me to recognize them while I was retriggering them.

Comment 6

[(inactive) Jim Chen [:jchen] [:darchons]]

Could be due to Bug 680190. Locally, the best way to check for this bug is to look for two identical "org.mozilla.fennec" processes in 'adb shell ps', because the hang happens after fork() and before exec().

Comment 7

[Phil Ringnalda (:philor)]

http://tbpl.allizom.org/php/getParsedLog.php?id=6122336

Comment 8

[John O'Duinn [:joduinn] (please use "needinfo?" flag)]

(In reply to Jim Chen [:jchen] (mobile intern :) from comment #6)
> Could be due to Bug 680190. Locally, the best way to check for this bug is
> to look for two identical "org.mozilla.fennec" processes in 'adb shell ps',
> because the hang happens after fork() and before exec().

bug#680190 was fixed on 31aug, so the big question is "are we still seeing this problem since 31aug?".

Comment 9

[Phil Ringnalda (:philor)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1315836932.1315839581.24663.gz

Comment 10

[Phil Ringnalda (:philor)]

Oh, no wonder this seems to be both common, and rare: this isn't the failure, or a symptom of the failure, but a separate failure of the mobile Talos harness to recognize that a Talos failure happened and deal appropriately, and because the Tinderbox errorparser highlighted this line because of an unfortunate regex, for people who didn't know to first look at the TinderboxPrints for Talos errors, it looked like this was the problem.

If you look at the full log for that last tinderbox log in comment 9 while it's still around, you'll see that it's actually a bug 675297 "timeout exceeded" Talos error, but bcontroller.py fails to realize what happened. If you look at https://tbpl.mozilla.org/php/getParsedLog.php?id=6599245&tree=Mozilla-Inbound you'll see that since the buildbot-based tbpl errorparser doesn't think this is a highlightable error, it doesn't put it in the summary popup, so someone looking at tbpl is going to look at the TinderboxPrinted "initialization timed out" rather than being distracted by this later harness error.

Comment 11

[[e-mail] (:bhearsum)]

I'm going to assume this isn't an issue anymore. Please re-open if it is.