Users are reporting crashes on start screen in latest beta |
||||||||||||||
Issue descriptionChrome Version: 69.0.3497.58 OS: Chrome So far seems to be eve specific, but we haven't been able to confirm. What steps will reproduce the problem? (1) Boot (2) Wait for a few seconds at the login screen What is the expected result? No crash What happens instead? Crash Not all users are affected, even on eve. Some people have reported that if they just allow it to sit for a while it will eventually recover and the machine will be useful again. This is possibly a manifestation of bug 865970, bug 866711, bug 876385, bug 865444, or is a reopen of bug 863261 . jdufault@ can I get you to drive investigation here?
,
Aug 27
,
Aug 27
Observations over the weekend: * The CL reverted in https://chromium-review.googlesource.com/1184542 is NOT the issue. Symptoms are different. * I encountered a similar set of issues on Dru over the weekend, but call stacks pointed at bug 865444 * Stacks from a crashing eve over the weekend pointed to bug 866711, and crashes in arc_networkd * Chrome OS crashes with stable signature "00000f7e-exit2-ui" are an expected side effect of an unexpected UI exit.
,
Aug 27
These are the list of crashes that we observed during our testing. These crashes were repro'd after login to user account but not while on start screen. 1. Issue 842272:- Browser crash: base::MessagePumpDefault::Run. Not 100% reproducible. Reproduced a couple of times immediately after sign-in to user account 2. Issue 877278:- Browser crash: base::MessagePumpDefault::Run. Device connected to external monitor and lid is closed. When the device lid is opened back (i.e. while exiting clamshell) browser crash is seen. 3. Issue 877177:- Browser crash: ash::WindowTreeHostManager::Shutdown. Not 100% reproducible. Browser crash was observed after sign-in to user account.
,
Aug 27
Looking at crash/, I suspect that this is from easy unlock (issue 866711). Trying to verify.
,
Aug 27
,
Aug 27
I haven't been able to repro, but I'm going to be OOO soon, passing to jhawkins@. I think this is either smart lock or arcd; the DRM crashes are unlikely because login shows up for a second or so, and the DRM issue implies that no rendering would happen. This could also be caused by display reconfiguration, ie, if the device is attached to a monitor, but I was not able to repro that way. The display reconfiguration bug has been fixed in a version after 69.0.3497.58 by afakhry@. Letting the device sit for a while and recover may mean that the device is reverting to the previous version of Chrome.
,
Aug 27
,
Aug 28
,
Aug 29
Here is stack trace from the debug ToT build: [9643:9643:0828/205824.278402:FATAL:device_sync_client_impl.cc(100)] Check failed: is_ready(). #0 0x56c73ed33b98 base::debug::StackTrace::StackTrace() #1 0x56c73ead64cc base::debug::StackTrace::StackTrace() #2 0x56c73eb1ee9a logging::LogMessage::~LogMessage() #3 0x56c73be5a69e chromeos::device_sync::DeviceSyncClientImpl::GetSyncedDevices() #4 0x56c73b72ca96 chromeos::EasyUnlockServiceRegular::GetUnlockKeys() #5 0x56c73b7320e3 chromeos::EasyUnlockServiceRegular::InitializeInternal() #6 0x56c73b7218cc chromeos::EasyUnlockService::InitializeOnAppManagerReady() #7 0x56c734f6adef _ZN4base8internal13FunctorTraitsIMN3net16HostResolverImpl8ProcTaskEFvvEvE6InvokeIS6_NS_7WeakPtrIS4_EEJEEEvT_OT0_DpOT1_ #8 0x56c73b728aca _ZN4base8internal12InvokeHelperILb1EvE8MakeItSoIRKMN8chromeos17EasyUnlockServiceEFvvERKNS_7WeakPtrIS5_EEJEEEvOT_OT0_DpOT1_ #9 0x56c73b728a60 _ZN4base8internal7InvokerINS0_9BindStateIMN8chromeos17EasyUnlockServiceEFvvEJNS_7WeakPtrIS4_EEEEEFvvEE7RunImplIRKS6_RKNSt3__15tupleIJS8_EEEJLm0EEEEvOT_OT0_NSF_16integer_sequenceImJXspT1_EEEE #10 0x56c73b72899c _ZN4base8internal7InvokerINS0_9BindStateIMN8chromeos17EasyUnlockServiceEFvvEJNS_7WeakPtrIS4_EEEEEFvvEE3RunEPNS0_13BindStateBaseE #11 0x56c734fed05c _ZNO4base12OnceCallbackIFvvEE3RunEv #12 0x56c73ed7e1a8 base::debug::TaskAnnotator::RunTask() #13 0x56c73eb39136 base::MessageLoop::RunTask() #14 0x56c73eb394de base::MessageLoop::DeferOrRunPendingTask() #15 0x56c73eb399d9 base::MessageLoop::DoWork() #16 0x56c73ed73a35 base::MessagePumpLibevent::Run() #17 0x56c73eb3882b base::MessageLoop::Run() #18 0x56c73ebbaaed base::RunLoop::Run() #19 0x56c73dd77fa3 ChromeBrowserMainParts::MainMessageLoopRun() #20 0x56c73840d811 content::BrowserMainLoop::RunMainMessageLoopParts() #21 0x56c738415eb0 content::BrowserMainRunnerImpl::Run() #22 0x56c73840055b content::BrowserMain() #23 0x56c73dd40217 content::RunBrowserProcessMain() #24 0x56c73dd42b96 content::ContentMainRunnerImpl::Run() #25 0x56c73dd376bc content::ContentServiceManagerMainDelegate::RunEmbedderProcess() #26 0x56c73dd5a84a service_manager::Main() #27 0x56c73dd3dbb3 content::ContentMain() #28 0x56c734e611b8 ChromeMain #29 0x56c734e610a2 main #30 0x7f6879181736 __libc_start_main #31 0x56c734e60f49 _start
,
Aug 29
I've filed a different issue 878653 for the last stack trace.
,
Aug 30
I have easy unlock enabled as well on my eve and ran into crashes on update to 69 beta
,
Aug 30
FYI I checked with Urs also (who was the first user to report this issue to me) and confirmed that he is also an Easy Unlock user. Not sure if that is the service to blame but the 2 users I know who experienced crashes are both users of EasyUnlock.
,
Aug 30
Kyle, can you look at the stack in comment 10?
,
Aug 30
Stack in comment #10 is issue 878653 .
,
Aug 31
Kyle fixed the issue in comment 10. We will need logs or crash IDs to move forward on this issue.
,
Aug 31
The stack trace in comment #10 isn't relevant to this bug. This bug is about a crash that users are reporting in beta channel, but the stack trace in comment #10 is a DCHECK() which is compiled out for release builds (beta channel is a release build).
,
Aug 31
Right, I'm not referring to that crash. I'm saying any outstanding crash issues on login screen need more information in order to proceed.
,
Sep 1
Caroline Samsung Chromebook Pro is exactly experiencing the issue reported by abodenha@chromium.org, Aug 27 (5 days ago) after applying Version 69.0.3497.73 (Official Build) beta (64-bit). Its so bad I am expecting my SCBP to brick after a cold boot anytime soon. I've used CB nearly every day since 2013 and have not experienced boot up issues as severe before. Is there remediation we can take such as power wash or fresh application of firmware? There is no attachment because the machine doe not boot, it simply hangs requires at least 3 attempts to boot up successfully.
,
Sep 2
Feedback sent 9:26 pm EDT Issue 878002 - Caroline 69.0.3497.73 beta Shut down > restart > grey screen (not black) > power light on alt + volup + x one time
,
Sep 2
SHut down and restart again > waited 51 seconds > screen went black > splash screen and sign in screen. Reproduced 3 times. Video https://photos.app.goo.gl/LLZPX3ahtsN7PAHs5
,
Sep 2
Re: Comment 21 dympeh@ are you able to quickly login before Chrome crashes? Some users have been able to do so as a temporary workaround. Also, if you can file feedback in case you are able to login, that will be great. Please reference puneetster@ or abodenha@ in the note.
,
Sep 2
Re: comment 22 @puneetster, I can sign in everytime once the sign in screen shows up. There are no reported crashes in chrome://crashes. Sent feedback again at 6:53 am EDT. "Feedback for issue 778002 comment 23"
,
Sep 2
Sorry, forgot to reference you and abodenha@. This seems a different issue than the actual crashing at the start screen or the bootloop some are reporting.
,
Sep 2
6:45 pm EDT feedback sent. Issue 878002 feedback from the sign in screen reference puneetster@ , abodenha@ Issues has changed from grey screen > wait 51 seconds > black screen > splash screen > sign in. Now it is a black screen from the beginning > wait for a long time > splash screen > sign in. 1. shut down via status area shut down icon > close lid 2. open lid > power light comes on on the side of Caroline > black screen After signing in, there's no "Chrome didn't shut down properly" error message and everything works fine.
,
Sep 3
Confirming sequence reported in comment 25 by @dymp... after about 3 to 5 boot attempts on Caroline. Turned off smartlock (I wasn't using it). Now using only password. Will report if this makes a difference.
,
Sep 4
By observation of Caroline. There are two issues. Getting to the SCBP sign-on screen requires waiting about 20 seconds to interrupt the boot sequence with a power-off/power-on attempt that works to instantiate the sign-on screen (otehr wise it stays black but clearly on). Noticeably after inputting password, the screen again goes black for 1-3 seconds while Caroline apparently pauses to work past a boot issue (a pattern like booting up a Windows machine or slow Macbook). Turning off smartlock made no difference. I have never connected an external monitor. I do use a BT mouse not track-pad.
,
Sep 4
,
Sep 5
+apronin
,
Sep 5
,
Sep 6
Chromebook Pixel (2013) Link was to have stopped receiving updates. But gained updates one or two updates since the notice (This device will no longer receive the latest software updates. ...). After receiving Version 69.0.3497.73 (Official Build) beta (64-bit). The 2013 Pixel boots fine to splash screen (fast as usual). After accepting pwd (user name is defaulted) the sign-in screen turns black and noticeably pauses before completing the sign-in process. I also noticed that after the sign-in Link does not have access to gmail with pressing the blue sign-in button to gmail. I'm pretty sure this the 2013 Chromebook automatically normally signed into gmail before receiving Version 69.0.3497.73. I don't expect Link to receive any updates (as updates should have stopped). I am just reporting my Link observation tied to Caroline start screen crashes on latest beta.
,
Sep 10
Initially reported issues was identified as fixed, closing this bug as resolved. If the issue persists, please open a bug or reopen with details. Thank you.
,
Sep 10
[Auto-generated comment by a script] We noticed that this issue is targeted for M-69; it appears the fix may have landed after branch point, meaning a merge might be required. Please confirm if a merge is required here - if so add Merge-Request-69 label, otherwise remove Merge-TBD label. Thanks.
,
Oct 23
|
||||||||||||||
►
Sign in to add a comment |
||||||||||||||
Comment 1 by abodenha@chromium.org
, Aug 27