Intermittent robocop failures from "ValueError: invalid literal for int() with base 10: '(1'"/'shortMsg=Process'

RESOLVED WORKSFORME

Status

defect
RESOLVED WORKSFORME
7 years ago
5 years ago

People

(Reporter: philor, Unassigned)

Tracking

({intermittent-failure})

Trunk
ARM
Android
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(1 attachment)

https://tbpl.mozilla.org/php/getParsedLog.php?id=16500022&tree=Mozilla-Inbound
Android no-ionmonkey Tegra 250 mozilla-inbound opt test robocop on 2012-10-26 12:53:39 PDT for push 2373c969584c
slave: tegra-048

22 INFO TEST-PASS | testAllPagesTab | Context menu has Share option - Share
INFO | runtests.py | Received unexpected exception while running application
Traceback (most recent call last):
  File "/builds/tegra-048/test/build/tests/mochitest/runtests.py", line 731, in runTests
    timeout = timeout)
  File "/builds/tegra-048/test/build/tests/mochitest/automation.py", line 1050, in runApp
    status = self.waitForFinish(proc, utilityPath, timeout, maxTime, startTime, debuggerInfo, symbolsPath)
  File "/builds/tegra-048/test/build/tests/mochitest/remoteautomation.py", line 66, in waitForFinish
    status = proc.wait(timeout = maxTime)
  File "/builds/tegra-048/test/build/tests/mochitest/remoteautomation.py", line 202, in wait
    while (self.dm.processExist(self.procName)):
  File "/builds/tegra-048/test/build/tests/mochitest/devicemanager.py", line 175, in processExist
    procList = self.getProcessList()
  File "/builds/tegra-048/test/build/tests/mochitest/devicemanagerSUT.py", line 458, in getProcessList
    processTuples += [[int(pidproc[1]), pidproc[2], int(pidproc[0])]]
ValueError: invalid literal for int() with base 10: '(1'
Looks like this is caused by the change for bug 799507: TypeError exception when running remote reftest.
https://tbpl.mozilla.org/php/getParsedLog.php?id=16740810&tree=Mozilla-Inbound

ValueError: invalid literal for int() with base 10: 'shortMsg=Process'
https://tbpl.mozilla.org/php/getParsedLog.php?id=16743406&tree=Firefox
Summary: Intermittent robocop failures from "ValueError: invalid literal for int() with base 10: '(1'" → Intermittent robocop failures from "ValueError: invalid literal for int() with base 10: '(1'"/'shortMsg=Process'
At first glance, these errors suggest that sutAgent's ps command is sometimes returning "(1" for either the uid or pid -- I don't see how that is possible. Any ideas?
(In reply to Geoff Brown [:gbrown] from comment #8)
> At first glance, these errors suggest that sutAgent's ps command is
> sometimes returning "(1" for either the uid or pid -- I don't see how that
> is possible. Any ideas?

No, it sounds very strange. I think we should have more verbose debugging for this case, so we can see more clearly what's happening here. I was thinking about this on Sunday and have a tentative patch.
The one annoying thing about this patch is that we'll still be seeing the same error, but people will probably file it somewhere else because it will have a different error traceback/debugging info. Not sure how to deal with that.
Attachment #678528 - Flags: review?(gbrown)
That's easy to deal with:

+ print "This is bug 805969, copy-paste the following\n"
Comment on attachment 678528 [details] [diff] [review]
Be more verbose about problems parsing process list

Review of attachment 678528 [details] [diff] [review]:
-----------------------------------------------------------------

I like philor's idea; r+ with the addition of something like that.
Attachment #678528 - Flags: review?(gbrown) → review+
Pushed: https://github.com/mozilla/mozbase/commit/ed7d82c345c166f641e3d67977532b31dc984949

I added the statement "ERROR: Unable to parse process list (bug 805969)" to the output. Hopefully that'll be good enough to link up with this.

I also added a unit test to make sure this was working as expected. :)

Note that we'll need to merge the latest version of mozdevice into m-c to see these changes.
These errors are showing up in bug 813858 now.
Whiteboard: [orange]
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.