Closed Bug 917517 Opened 7 years ago Closed 5 years ago

Intermittent browser_thumbnails_background_crash.js | dumpID is present and not an empty string

Categories

(Toolkit :: General, defect)

x86_64
Linux
defect
Not set

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: RyanVM, Assigned: adw)

Details

(Keywords: intermittent-failure, Whiteboard: p=13)

Attachments

(1 file)

https://tbpl.mozilla.org/php/getParsedLog.php?id=27973810&tree=Mozilla-Inbound

Rev3 Fedora 12x64 mozilla-inbound debug test mochitest-browser-chrome on 2013-09-17 06:16:58 PDT for push a0fcc3eb892f
slave: talos-r3-fed64-062

07:40:37     INFO -  TEST-PASS | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | Thumbnail should be cached after capture: /tmp/tmp6ZKh5x/thumbnails/e82891364309427ebb2a6931429ff068.png
07:40:37     INFO -  TEST-INFO | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | Crashing the thumbnail content process.
07:40:37     INFO -  XPCOM_MEM_BLOAT_LOG: /tmp/tmp6ZKh5x/runtests_leaks.log
07:40:37     INFO -  Writing to log: /tmp/tmp6ZKh5x/runtests_leaks_tab_pid2795.log
07:40:43     INFO -  ###!!! [Parent][AsyncChannel] Error: Channel error: cannot send/recv
07:40:43     INFO -  [Parent 2405] WARNING: No docshells for remote frames!: file ../../../../content/base/src/nsFrameLoader.cpp, line 571
07:40:43     INFO -  [Parent 2405] WARNING: No docshells for remote frames!: file ../../../../content/base/src/nsFrameLoader.cpp, line 571
07:40:43     INFO -  TEST-PASS | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | Thumbnail should not have been saved due to the crash: /tmp/tmp6ZKh5x/thumbnails/80216e90ce8a2cf81b2fccee8d0925de.png
07:40:43     INFO -  [Parent 2405] WARNING: RemoveObserver() called for unregistered observer: file ../../hal/Hal.cpp, line 205
07:40:43     INFO -  [Parent 2405] WARNING: RemoveObserver() called for unregistered observer: file ../../hal/Hal.cpp, line 205
07:40:43     INFO -  [Parent 2405] WARNING: RemoveObserver() called for unregistered observer: file ../../hal/Hal.cpp, line 205
07:40:43     INFO -  [Parent 2405] WARNING: RemoveObserver() called for unregistered observer: file ../../hal/Hal.cpp, line 205
07:40:43     INFO -  [Parent 2405] WARNING: RemoveObserver() called for unregistered observer: file ../../hal/Hal.cpp, line 205
07:40:43     INFO -  TEST-PASS | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | Received correct observer topic.
07:40:43     INFO -  TEST-PASS | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | Subject implements nsIPropertyBag2.
07:40:43  WARNING -  TEST-UNEXPECTED-FAIL | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | dumpID is present and not an empty string
07:40:44     INFO -  Stack trace:
07:40:44     INFO -      JS frame :: chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js :: crashObserver :: line 140
07:40:44     INFO -      native frame :: <unknown filename> :: <TOP_LEVEL> :: line 0
07:40:44     INFO -  TEST-INFO | chrome://mochitests/content/browser/toolkit/components/thumbnails/test/browser_thumbnails_background_crash.js | Console message: [JavaScript Error: "BackgroundThumbnails remote process crashed - recovering" {file: "resource://gre/modules/BackgroundPageThumbs.jsm" line: 195}]
Summary: Intermittent browser_thumbnails_background.js | Captured URL should be currently expected URL (i.e., capture() callbacks should be called in the correct order) | Thumbnail should be cached after capture → Intermittent browser_thumbnails_background_crash.js | dumpID is present and not an empty string
Check out this comment where the dumpID property is added to the property bag:

http://mxr.mozilla.org/mozilla-central/source/dom/ipc/ContentParent.cpp#1246

Either that's happening, or the crash reporter can't get the ID from the minidump file:

http://mxr.mozilla.org/mozilla-central/source/dom/ipc/CrashReporterParent.cpp#94

Related to the latter possibility, bug 879772 was also an orange where minidumps weren't being generated as they should.

Either way, it doesn't seem like checking the dumpID is necessary.  Before it checks the dumpID, the test has ensured that a crash has happened and that it's abnormal, and I think that's all it needs to do.  But maybe I'm missing something?
Attachment #812381 - Flags: review?(mhammond)
Comment on attachment 812381 [details] [diff] [review]
don't check dumpID

Review of attachment 812381 [details] [diff] [review]:
-----------------------------------------------------------------

I'm fine with that if it solves the problem.  However, http://mxr.mozilla.org/mozilla-central/source/dom/ipc/tests/ does the same thing and doesn't seem to strike this.  Also, if it turns out we "miss" a dumpId, failing to remove it will cause the test runner to complain at the end of the test about the crash - so watch out for that.
Attachment #812381 - Flags: review?(mhammond) → review+
Oh, hmm, thanks.  It would be good to get an explanation for why test_process_error.xul doesn't have a similar failure.  I'll look some more.

Missing a dump ID and the test runner complaining about the crash -- I wonder whether that's what this is about, at the end of the failing logs:

> 10:53:47  WARNING -  PROCESS-CRASH | Main app process exited normally | application crashed [Unknown top frame]
> 10:53:47     INFO -  Crash dump filename: /tmp/tmpfDED13/minidumps/790c9b0d-05ff-8835-402013d1-74de3dac.dmp
> 10:53:47     INFO -  stderr from minidump_stackwalk:
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump_processor.cc:264: INFO: Processing minidump in file /tmp/tmpfDED13/minidumps/790c9b0d-05ff-8835-402013d1-74de3dac.dmp
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump.cc:3815: INFO: Minidump opened minidump /tmp/tmpfDED13/minidumps/790c9b0d-05ff-8835-402013d1-74de3dac.dmp
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump.cc:4086: ERROR: ReadBytes: read 0/32
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump.cc:3834: ERROR: Minidump cannot read header
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump_processor.cc:268: ERROR: Minidump /tmp/tmpfDED13/minidumps/790c9b0d-05ff-8835-402013d1-74de3dac.dmp could not be read
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump.cc:3787: INFO: Minidump closing minidump
> 10:53:47     INFO -  2013-09-24 10:53:47: minidump_stackwalk.cc:529: ERROR: MinidumpProcessor::Process failed
> 10:53:47     INFO -  minidump_stackwalk exited with return code 1
Assignee: nobody → adw
Status: NEW → ASSIGNED
Many try runs with logging later, can't reproduce... :-/

https://tbpl.mozilla.org/?tree=Try&rev=39a1016ad6ef
Whiteboard: [triage]
Whiteboard: [triage]
Whiteboard: p=0
No longer blocks: fxdesktopbacklog
Flags: firefox-backlog+
Whiteboard: p=0 → p=13
Inactive; closing (see bug 1180138).
Status: ASSIGNED → RESOLVED
Closed: 5 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.