New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 808979 link

Starred by 2 users

Issue metadata

Status: WontFix
Owner:
Closed: Mar 2018
Cc:
EstimatedDays: ----
NextAction: ----
OS: Android
Pri: 1
Type: ----

Blocked on:
issue 813832



Sign in to add a comment

memory.long_running_dual_browser_test failing on internal.client.clank/integration-android-low-end-phone

Project Member Reported by sheriff-...@appspot.gserviceaccount.com, Feb 5 2018

Issue description

Filed by sheriff-o-matic@appspot.gserviceaccount.com on behalf of gsennton@google.com

memory.long_running_dual_browser_test failing on internal.client.clank/integration-android-low-end-phone

Builders failed on: 
- integration-android-low-end-phone: 
  https://uberchromegw.corp.google.com/i/internal.client.clank/builders/integration-android-low-end-phone


Started failing at build nr #4344:
https://uberchromegw.corp.google.com/i/internal.client.clank/builders/integration-android-low-end-phone/builds/4344

The failing test is
google_tip_on_100_bill

 

Comment 1 by ctzsm@chromium.org, Feb 14 2018

This is still happening, with less frequency though. Could someone please take a look? Thanks!
Owner: perezju@chromium.org
Status: Assigned (was: Available)
Status: WontFix (was: Assigned)
As of https://uberchromegw.corp.google.com/i/internal.client.clank/builders/integration-android-low-end-phone/builds/4393 the bot has been mostly green.

Only three failures, each with a single failing story, and always a different one:
#4394 google
#4401 amazon
#4408 wikipedia

So no consistent failure to investigate.

A single story failing out of 1200 tests run on this benchmark is an acceptable ratio for this bot.

Comment 4 by donnd@google.com, Feb 20 2018

Labels: OS-Android
Status: Untriaged (was: WontFix)
Looks like this is happening more frequently, failing every day now: https://uberchromegw.corp.google.com/i/internal.client.clank/builders/integration-android-low-end-phone

Is it possible to just add some kind of retry-on-failure so this issue doesn't keep coming up on SOM?
This looks like a new consistent failure. Unrelated to the previous ones.

The failing story is http_en_m_wikipedia_org_wiki_Science, with the browser actually crashing:

	signal 11 (SIGSEGV), code 1, fault addr 0xe00cf81e in tid 4194 (Chrome_IOThread)
	pid: 4152, tid: 4194, name: Chrome_IOThread  >>> com.google.android.apps.chrome <<<
	signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0xe00cf81e
	     r0 00000000  r1 a2611a88  r2 00001000  r3 b9a6d50c
	     r4 b9a6d4d8  r5 00000000  r6 b9a6d4f4  r7 00000040
	     r8 a0db54b0  r9 e00cf81e  sl b9a6d50c  fp 00000000
	     ip 070cca8b  sp a0db54a0  lr 988f48a7  pc 988f498c
	
	Stack Trace:
	  RELADDR   FUNCTION                                                                                                                                                      FILE:LINE
	  0077698c  mojo::edk::PlatformChannelRecvmsg(mojo::edk::ScopedPlatformHandle const&, void*, unsigned int, base::circular_deque<mojo::edk::ScopedPlatformHandle>*, bool)  /b/build/slave/arm-builder-rel/build/src/mojo/edk/embedder/platform_channel_utils_posix.cc:182:17

Full logs attached. I'll also kick off a bisect, hopefully will help to narrow down a culprit. From the crash message sounds like this could be related to issue 734645?
wikipedia_crash_logs.txt
86.2 KB View Download
Cc: aluo@chromium.org
Actually, this is the long running test. Bisect won't work here. :-/
Labels: -Pri-2 Pri-1
Blockedon: 813832
Sounds very much related to issue 813832, blocking on that for now.
Status: Assigned (was: Untriaged)
Status: WontFix (was: Assigned)
Has been again mostly green since:
https://uberchromegw.corp.google.com/i/internal.client.clank/builders/integration-android-low-end-phone/builds/4463

If new failures occur please file a new bug.

Sign in to add a comment