Closed Bug 1340621 Opened 9 years ago Closed 8 years ago

Intermittent test_refresh_firefox.py TestFirefoxRefresh.testReset | IOError: Process has been unexpectedly closed (Exit code: 0) (Reason: Timed out waiting for connection on localhost:2828!)

Categories

(Toolkit :: Startup and Profile System, defect)

Product:
Toolkit
Component:
Startup and Profile System
Platform:
Unspecified
Windows
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla58
Tracking Flags:
Tracking Status
firefox57 --- fixed
firefox58 --- fixed

People

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

References

Details

(Keywords: intermittent-failure)

See Also: → 1274120
All the failing jobs show an assertion during this test: --DOCSHELL 0000009A26939000 == 0 [pid = 3292] [id = {4a8758c3-da6c-4c42-b817-c33b6062cd03}] [GPU 3660] WARNING: pipe error: 109: file c:/builds/moz2_slave/m-in-w64-d-0000000000000000000/build/src/ipc/chromium/src/chrome/common/ipc_channel_win.cc, line 346 Attempt to add a bool pref cache for preference 'dom.w3c_pointer_events.enabled' at address '000007FDE64D4A85'was made. However, a pref was already cached at this address. Assertion failure: false (Should not have an existing pref cache for this address), at c:/builds/moz2_slave/m-in-w64-d-0000000000000000000/build/src/modules/libpref/Preferences.cpp:266 #01: ???[C:\slave\test\build\application\firefox\xul.dll +0x52c2dc] #02: mozilla_dump_image[C:\slave\test\build\application\firefox\xul.dll +0x2dac69c] #03: mozilla_dump_image[C:\slave\test\build\application\firefox\xul.dll +0x2d661ac] #04: ???[C:\slave\test\build\application\firefox\xul.dll +0x4da897] #05: ???[C:\slave\test\build\application\firefox\xul.dll +0x4d9982] #06: ???[C:\slave\test\build\application\firefox\xul.dll +0x4d8719] #07: PrintJSStack[C:\slave\test\build\application\firefox\xul.dll +0x1228eb8] #08: DumpFrameArray[C:\slave\test\build\application\firefox\xul.dll +0x3c24a94] #09: DumpFrameArray[C:\slave\test\build\application\firefox\xul.dll +0x3c25b43] #10: DumpFrameArray[C:\slave\test\build\application\firefox\xul.dll +0x3c1e68c] #11: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3f06eab] #12: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3f7c116] #13: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3f7e988] #14: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3f7d53b] #15: ???[C:\slave\test\build\application\firefox\xul.dll +0x510d82] #16: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xf353db] #17: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xf354d3] #18: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xf557d4] #19: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x4380b58] #20: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x48ec569] #21: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x48effbd] #22: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x48f889c] #23: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x48ec62c] #24: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x48e4d74] #25: XRE_GetBootstrap[C:\slave\test\build\application\firefox\xul.dll +0x41b50c1] #26: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xf3615f] #27: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xf3553b] #28: ???[C:\slave\test\build\application\firefox\xul.dll +0x510cc1] #29: ???[C:\slave\test\build\application\firefox\xul.dll +0x510dd9] #30: ???[C:\slave\test\build\application\firefox\xul.dll +0x4deeb9] #31: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3fb76c6] #32: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3fb154d] #33: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3fae5e3] #34: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3fade14] #35: ???[C:\slave\test\build\application\firefox\firefox.exe +0x1996] #36: ???[C:\slave\test\build\application\firefox\firefox.exe +0x141e] #37: ???[C:\slave\test\build\application\firefox\firefox.exe +0x1d5b] #38: TargetNtUnmapViewOfSection[C:\slave\test\build\application\firefox\firefox.exe +0x470f5] #39: BaseThreadInitThunk[C:\windows\system32\KERNEL32.DLL +0x167e] #40: RtlUserThreadStart[C:\windows\SYSTEM32\ntdll.dll +0x1c3f1] [GPU 3660] WARNING: Shutting down GPU process early due to a crash!: file c:/builds/moz2_slave/m-in-w64-d-0000000000000000000/build/src/gfx/ipc/GPUParent.cpp, line 454 The assertion is pretty useless as long as bug 1348961 hasn't been implemented. But i wonder if that was causing a crash, which then triggered a shutdown of the GPU process.
Depends on: 1348961
Here some info about the shutdown which seems to happen here instead of a restart. Maybe this is related to a hang/crash on shutdown (see bug 1375242 and bug 1374043) 17:08:22 INFO - 1499299702330 Marionette DEBUG Received observer notification "xpcom-shutdown" 17:08:22 INFO - 1499299702330 Marionette DEBUG Received observer notification "xpcom-shutdown" 17:08:22 INFO - Unable to read VR Path Registry from C:\Users\cltbld.T-W864-IX-226.001\AppData\Local\openvr\openvrpaths.vrpath 17:08:22 INFO - [GPU 2192] WARNING: pipe error: 109: file c:/builds/moz2_slave/autoland-w64-00000000000000000/build/src/ipc/chromium/src/chrome/common/ipc_channel_win.cc, line 346 17:08:22 INFO - ###!!! [Child][MessageChannel::SendAndWait] Error: Channel error: cannot send/recv 17:11:27 ERROR - TEST-UNEXPECTED-ERROR | test_refresh_firefox.py TestFirefoxRefresh.testReset | IOError: Process has been unexpectedly closed (Exit code: 0) (Reason: Timed out waiting for connection on localhost:2828!) 17:11:27 Not sure why Marionette receives the `xpcom-shutdown` observer notification twice but this could be similar to bug 1254136, which I will have a look at soon. But it shouldn't cause this issue.
https://treeherder.mozilla.org/logviewer.html#?repo=autoland&job_id=112808022&lineNumber=63179 So after the restart I can see that Marionette gets enabled via the environment variable: > 18:19:07 INFO - 1499563147966 Marionette INFO Enabled via MOZ_MARIONETTE 1 But that's all. There is no "profile-after-change" observer notification fired by Firefox which sets-up further observers for Marionette to get the server socket started.
Component: Marionette → Startup and Profile System
Product: Testing → Toolkit
Version: Version 3 → unspecified
This is exclusively a Windows only test failure those days.
OS: Unspecified → Windows
Blocks: 1386590
Gabriele, that's another case when Breakpad crashes analyzing the minidump file: https://treeherder.mozilla.org/logviewer.html#?job_id=125280202&repo=mozilla-inbound&lineNumber=57801 I assume it's the same issue as bug 1391766?
Flags: needinfo?(gsvelto)
Yeah, it looks the same, these are the first two frames of the browser stack: 0 xul.dll!google_breakpad::ExceptionHandler::WriteMinidump(std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator<wchar_t> > const &,bool (*)(wchar_t const *,wchar_t const *,void *,_EXCEPTION_POINTERS *,MDRawAssertionInfo *,bool),void *,_MINIDUMP_TYPE) [exception_handler.cc:d97a8d7abdd6 : 764 + 0xb] eip = 0x58a314fc esp = 0x1a95f130 ebp = 0x1a95f210 ebx = 0x00000001 esi = 0x1a95f138 edi = 0x000006e0 eax = 0x1a95ee00 ecx = 0x00000000 edx = 0x000006e0 efl = 0x00000216 Found by: given as instruction pointer in context 1 xul.dll!CrashReporter::CreateMinidumpsAndPairInternal(void *,unsigned long,nsCString,nsCOMPtr<nsIFile>,nsIFile * *,std::function<void > &&,RefPtr<nsIThread> &&,bool) [nsExceptionHandler.cpp:d97a8d7abdd6 : 4115 + 0x14] eip = 0x58a2b91d esp = 0x1a95f218 ebp = 0x1a95f2cc Found by: call frame info As you can see we're in the same scenario, and as Ted noticed in bug 1391766 comment 7 ecx is null so this is crashing after failing to write a child process. BTW that code received a pretty heavy refactoring in bug 1360308 but I'm not sure if it's related since I seem to remember we already had these errors before that bug landed.
Flags: needinfo?(gsvelto)
The last reported Linux failures were wrongly starred and should have been filed as a new bug because those are actually crashes of Firefox as Treeherder clearly shows. But given that the patch on bug 1381748 has already been landed, I don't think we need a new bug anymore.
Depends on: 1314885
The failure here happens after an in_app restart. As such it should no longer happen once my patch on bug 1410366 got landed.
Depends on: 1410366
(In reply to Henrik Skupin (:whimboo) from comment #33) > The failure here happens after an in_app restart. As such it should no > longer happen once my patch on bug 1410366 got landed. It did land, and got uplifted, a week ago, and no OF bot update suggests this is now fixed! \o/
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Fantastic to hear. Completely missed to check that bug again, so thank you for doing it.
Assignee: nobody → hskupin
status-firefox57: --- → fixed
status-firefox58: --- → fixed
Target Milestone: --- → mozilla58
No longer depends on: 1314885, 1348961
You need to log in before you can comment on or make changes to this bug.