Issue metadata
Sign in to add a comment
|
After updating to 62.0.3202.18 beta, all tabs and all extensions crash immediately on startup
Reported by
mich...@michaelmarley.com,
Sep 15 2017
|
||||||||||||||||||||||
Issue descriptionUserAgent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:56.0) Gecko/20100101 Firefox/56.0 Steps to reproduce the problem: 1. Install Chrome Beta 62.0.3202.18 on Linux (not sure how hardware specific this is; I haven't tried it on any other systems yet because it doesn't seem possible to downgrade. 2. Launch Chrome Beta 3. Witness all tabs and all extensions crashing What is the expected behavior? The browser should start and run normally without crashing. What went wrong? All tabs crash with the Sad Tab error and all extensions crash and display a system notification. In the stdout/stderr, a number of instances of the following can be seen: [4143:4177:0915/065332.776216:ERROR:zygote_communication_linux.cc(146)] Did not receive ping from zygote child [5:5:0915/065332.776334:ERROR:zygote_linux.cc(627)] Zygote could not fork: process_type renderer numfds 6 child_pid -1 In dmesg, the following is repeated: [ 933.374608] traps: chrome[4272] trap int3 ip:5591bbbe3f5c sp:7ffdb1d5d210 error:0 Crashed report ID: No, the "chrome://crashes" page crashes How much crashed? Whole browser Is it a problem with a plugin? No Did this work before? Yes The last beta immediately before 62.0.3202.18, which I believe was still 61. Chrome version: 62.0.3202.18 Channel: beta OS Version: 4.13.2 Flash Version:
,
Sep 15 2017
Since it didn't show up anywhere in the initial report, I am running Kubuntu 17.10 x86_64 with kernel 4.13.2. The graphics card is an NVIDIA card with driver version 384.69, if that matters.
,
Sep 15 2017
michael@, Thank you for the report. Are you seeing any crash ids# from chrome://crashes?
,
Sep 15 2017
I apologize, but I cannot retrieve any info from "chrome://crashes" because the page will not load; it crashes instead.
,
Sep 15 2017
Thank you for providing more feedback. Adding requester "manoranjanr@chromium.org" to the cc list and removing "Needs-Feedback" label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Sep 20 2017
Can you get the crash IDs when you run with --no-sandbox?
,
Sep 20 2017
I tried enabling the crash reporting, restarting without "--no-sandbox" to reproduce the crashes, and restarting again with "--no-sandbox" to look at the crash reports, but no crash reports were shown. Sorry.
,
Oct 10 2017
@ michael-- Could you please confirm if you can still see this issue in latest beta M62 #62.0.3202.45 and update us with your observations. Thanks!
,
Oct 10 2017
The problem is still going on with that version and I still can't get any data from "chrome://crashes". I have, however, noticed that, at least as far as my systems go, it only seems to crash on the PCs with multiple monitors. I use two such desktops, one with NVIDIA graphics and one with Intel graphics and it crashes on both of those. I also have tried on two laptops and an HTPC that only have one monitor each (one has an Intel GPU, one has a modern NVIDIA GPU, and one has a legacy NVIDIA GPU) and it doesn't crash on any of those. (All the systems are running the same OS._
,
Oct 10 2017
Thank you for providing more feedback. Adding requester "hdodda@chromium.org" to the cc list and removing "Needs-Feedback" label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Oct 10 2017
Reporter, 1. look for the latest crash dump file in a directory like ~/.config/google-chrome/Crashpad/reports/ 2. or try bisecting the bug https://www.chromium.org/developers/bisect-builds-py
,
Oct 10 2017
There is no such directory on my system. There is a "Crash Reports" directory, but that directory is empty. I don't have a computer capable of compiling Chrome in a timely fashion, so bisecting would take an inordinately large amount of time. On 2017-10-10 11:39, wox… via monorail wrote:
,
Oct 10 2017
The bisect-builds script doesn't require building Chromium. It pulls prebuilt binaries for you to test. All that's required is: $ curl -s --basic -n "https://chromium.googlesource.com/chromium/src/+/master/tools/bisect-builds.py?format=TEXT" | base64 -d > bisect-builds.py $ python bisect-builds.py -a linux64 -g 400000
,
Oct 12 2017
,
Oct 18 2017
michael@, Gentle ping as per c#13.
,
Oct 19 2017
Running Chrome STABLE 62.0.3202.62 in 64bit openSUSE Leap 42.3, Plasma 5.11.1, KDE Frameworks 5.38.0, Qt 5.9.2, kernel 4.4.90-28-default, NVIDIA 384.90. Chrome 61 stable was working in Plasma 5.11, Chrome 62 stable does not work in neither Plasma 5.11.1 nor Plasma 5.11 (snapper rollback). Starting Chrome from command line: [11942:11976:1019/115303.954857:ERROR:zygote_communication_linux.cc(146)] Did not receive ping from zygote child [5:5:1019/115303.954978:ERROR:zygote_linux.cc(627)] Zygote could not fork: process_type renderer numfds 6 child_pid -1 Chrome runs, does not crash, but is not working/usable at all - no pages can be displayed including internal ones (chrome:// ...), all extensions are disabled/not started
,
Oct 19 2017
I can report the same behavior than Comment 17
,
Oct 23 2017
Same symptoms as others, specifically Chrome V62.x with blank tabs and extensions which crash.
Environment is
Video:Dell Radeon R9 360 OEM
Intel HD Graphics 530 (SkyLake GT2)
openSuSE Leap 42.2. Have tried gnome and kde.
kDE Plasma: 5.8.6, kernel Version 4.4.90-18.32-default
KDE frameworks 5.26.0
QT version 5.6.1
gnome 3.20.2
Results of bisect-builds.py
Revision 460256 (59.0.3055.0) WORKS
486049 (61.0.3156.0) WORKS
498920 (Can't see version) DOES NOT WORK
,
Oct 23 2017
RE:Comment 19 Here is a more granular look at what worked and did not.
,
Oct 23 2017
You need to keep running the script until it comes down to the minimal range and prints a big chunk of text and a long changelog URL.
,
Oct 25 2017
|
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by mich...@michaelmarley.com
, Sep 15 2017