V 6.1 | Sporadic Crashes
-
crashes sporadically, no pattern observed but caught a crash sequence mesg
Vivaldi has been crashing (just completely instantaneously terminates). Did not happen until last month or so. Keeps getting more frequent.
If there is a combination of tabs that crashes vivaldi, it will keep crashing until you change one of the tabs. Then it runs again.Started in a terminal in one of these "crash" situations. See below:
om@mydesktop: ~ vivaldi [18832:18832:0429/211444.410021:ERROR:chrome_browser_cloud_management_controller.cc(162)] Cloud management controller initialization aborted as CBCM is not enabled. [18832:18854:0429/211444.763162:ERROR:nss_util.cc(357)] After loading Root Certs, loaded==false: NSS error code: -8018 [18832:18855:0429/211445.430161:ERROR:object_proxy.cc(590)] Failed to call method: org.freedesktop.DBus.StartServiceByName: object_path= /org/freedesktop/DBus: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. [0429/211448.376517:ERROR:elf_dynamic_array_reader.h(64)] tag not found [0429/211448.376604:ERROR:elf_dynamic_array_reader.h(64)] tag not found Segmentation fault (core dumped) tom@mydesktop: ~ code_text
Using vivaldi version 6.1.3035.111 in linux.
Beyond my knowledge sphere to solve this. Any ideas?
tom kosvic--
ModEdit: Title
-
Hi,
--
Please,
On each report add:- Vivaldi Version: |
- Since when happens: |
- OS / Version / DE |
--
Vivaldi 6.1?
Is that right?Any reason for that?
--
Start with the Basic Desktop Troubleshooting Steps.
Extras
- Clean Site Data
- Clean Service Workers
vivaldi://serviceworker-internals/
- Reset Flags
vivaldi://flags
[If personally enabled]
--
Also,
Some useful links:Forums Community Official Tutorials Official Help Forum Categories Modding Vivaldi Vivaldi Features Vivaldi Help Forum Markdown Panels • Engines Vivaldi Tutorials Issues Feature Requests ¿? Menus • Guides • FAQ Vivaldi How To Bug Reports
--
Avoid Data loss
Follow the Backup | Reset links below
Vivaldi Backup | Reset + Extra Steps
-
@tckosvic ,
gee, did you not see it, I put the version in a line of text under the mesg text. Yes 6.1.
Using openSUSE leap 15.5tom kosvic
-
I Saw the version, but currently 6.1 is very old.
Any reason?
Old OS support?You should Backup and Upgrade to the latest 6.7
-
@Boefjim If the browser crashes it is always the browser that is at fault, not the web site. I have the same problem and logging out of Google works around it, but that's hardly a long-term solution.
I reported this through Vivaldi's crash reporting and it was closed as not reproducible in less than an hour.
-
@flodin If you are referring to VB-105797 (it was filed using your email address), that report was filed using Vivaldi 6.2 for an issue using Vivaldi 6.2 , not the then-current Vivaldi 6.6 (and now v6.7).
The tester who sent you the email, advised you to make sure you was using the newest version; the tester was also unable to reproduce in the then-current version. (re updating on your OS, as that is done by an independent maintainer that is apparently no longer doing so, you should just have received an email about how to switch to Flatpak.)
Generally, we don't go back 6+ months to analyze crashes (6.2 was released in September; we actually tried earlier today for another report, but no longer had the necessary data); if they are not happening in the most recent version (or the one before, currently 6.6), that indicates that the issue have probably been fixed months ago.
Crashes may be the "fault" of the application, but there are different kinds of crashes. Two of them are OOM (out of memory, can no longer work) and CHECK-asserts ("this state is already sufficiently broken that we will crash sooner rather than later, crash the browser here so that it is easy to determine why it crashed", which may be triggered by websites or extensions breaking stuff).
Regarding the others reporting crashes in this thread, I could not find any report using your names or emails, which also means we could not have analyzed what caused your problems (and the volunteers who tried to help you in this thread wasn't able to reproduce when they tried, and thus could not report a separate issue, either).
-
@yngve I interpreted the email differently, but it's probably not productive to quibble over semantics. However, I do want to say that I think your policy to not reproduce crashes in 6 month old versions is problematic. Few Linux distributions have release cycles that are shorter than that, and a crash bug is often an opportunity for exploitation by malicious actors. In this case, it is an out-of-bounds memory access which to me indicates something exploitable. In my opinion it's kind of a social responsibility as a developer to take such bugs seriously even if they happen in a six months old version. I did note in my initial crash report that 6.2 is the most recent version available to me on my Linux distribution and that I am unable to upgrade to a newer version than that.
Vivaldi is the only browser I have used in the past decade where the entire chrome has crashed, which is remarkable and surprising since Chromium which it is based on has process-based isolation that will contain crashes to the browser tab in which they occur.
This is aggravating not just because you lose the work you have done in every tab, but because Vivaldi preserves the crashing tab in the session that it reloads on next start. So the next time you start the browser it immediately dies again and again, and to be able to start at all you have to delete your entire profile which means losing all your session cookies. It's a poor user experience and it took me quite a bit of web searching (in a web browser of a different brand) before I was able to figure out how to get up and running again - definitely not something my dad would be able to figure out.
For these reasons, I urge you to reevaluate your stance that it's OK for web sites to trigger CHECK asserts that crash the entire chrome. Such asserts certainly have their place, but there are steps you can take to limit their impact when they fail.
[Note: in parallel to this discussion I have received additional emails from the Vivaldi Team suggesting I use Flatpak to install a newer version of Vivaldi instead; I will try this as soon as I have the opportunity.]
-
@flodin Vivaldi is not a Linux distro, and even Linux distros release security patches every few weeks (as does Windows).
Vivaldi's major version cycle is normally 2 months, the same as Chromium Extended Stable, the minor version cycle is usually once a week, and usually contains security patches (this past week we've released three (3) releases, two of which were for 0-day problems in Chromium). Analyzing crashers usually only makes sense for the two last major versions (presently 6.6. and 6.7).
We have to assume and depend on users keeping updated to the most recent version. In your case you relied on a non-Vivaldi-managed update channel, which (for some reason) stopped working about 6 months ago.
If a user does not update, they are going to be vulnerable to to quite a few security problems.
As mentioned, something that was broken 8 months ago may have been fixed 6 months ago, or last month. The chromium team alone adds thousands of updates every week, any of which may fix one issue, or cause another new one.
As mentioned, the tester attempted to reproduce your issue, and failed, which is why you were advised to update to the newest version to see if you could still reproduce the problem (if they had succeeded we would have gotten a crashdump that we could have analyzed). That failure to reproduce the problem suggests that either the problem no longer existed, or there were missing information about components of your configuration that impacted the case (e.g settings or Extensions).
Regarding the CHECKs: just about ALL of them (10000+) were placed there by the Chromium developers, not Vivaldi, and many of them will crash the whole browser because it is more secure than allowing a bad state to possibly compromise the security of the browser, and some of the others may crash Vivaldi's main process (which is a web app).
-
I have been encountering same crash problem. I posted thread a couple of weeks ago "V 6.1 | Sporadic Crashes" reporting this. Commentators suggested my problem was caused by using older version. I updated to 6.7 linux and problem persists sporadically.
I am also able to save vivaldi by deleting latest tab. Until i can delete that tab, vivaldi keep crashing.Unlike others though i cannot relate crashes to specific websites. Best correlation I have seen is that it is always the third tab that causes the crash but not always. When it happens it appears to be on the loading of the third tab.
hope, this helps. tom kosvic
-
@tckosvic I can't find any recent crash bug reports from you. We can't really analyze such issues without such data.
You might, however, want to make sure you update to the 6.7 update that was released a few hours ago, which also includes some crash fixes.
BTW, just because two problems involve the same event (e.g. crashes), that does not mean they are caused by the same thing. If the event happens when doing the exact same thing, yes, then the reason may well be the same, but only by analyzing crashdumps can we really be sure about that.
One thing to keep in mind about a pattern that does not always match, is to determine what the difference in the two instances are (e.g. does some of the pages involve authentication dialogs? Other special functionality?).
-
@yngve , In my initial thread reporting sporadic crashes that is referenced in my vivaldi forum --linux thread text as: "V 6.1 | Sporadic Crashes", I presented error messages when loading vivaldi from a terminal when it was crashing. Note, someone took it upon themselves to change my original title of my thread. Don't know who or why.
If the code included in this thread is not an official crash report, please point me how to file same.
tom kosvic
-
@tckosvic I don't think any of those lines are relevant to the problem, aside from the "segmentation fault" line (Chromium can be a bit vocal at times)
Filing bugs: See https://help.vivaldi.com/desktop/troubleshoot/reporting-a-bug-in-vivaldi/
Re title, the moderators have been known to modify subject titles so that they better describe what the post is about. A general example is all-caps (shouty) titles. In your case I guess the mods found your original subject line a little bit too long, with most of it instead belonging inside the text of the post.
-
Hi,
As @yngve mentioned, too long Titles are shortened.Original Titles are added to the Topic's Body as the first line and bold.
https://forum.vivaldi.net/topic/97420/v-6-1-sporadic-crashes
-
@Zalex108 , I try to make titles that luminate the problem so that people will look at the post. Too short meaningless not descriptive titles are worthless for getting a response.
If you want shorter titles, let that be known and reduce the allowable text space for titles instead of editing peoples work.
More importantly though, you found nothing useful in the terminal error messages I captured. How could I get some meaningful info if vivaldi just crashes and if I can get something where do I paste it? In the crashes they don't even run the "exit" dialogue when they occur.
Just trying to help solve a vivaldi problem I have observed. I am a very satisfied multi-year user and have not had these sporadic total crashes before. This is just as Boefjim has observed.
thanks, tom kosvic
-
no pattern observed but caught a crash sequence mesg
This doesn't bring anything useful to a Title.3 or more lines in a Title, is a waste of reading time aside of the fact that on modern times, many people Reads on a Mobile phone so it's also a waste of space on the screen.
This statement is valid as well for all the Topis around.
Choose a clear and concise title for the topicSame pattern as in an eMail subject.
--
On the other side,
You are mixing.The Terminal error it's in your original post, that which you left without answer.
The reading of a Terminal error, being V 6.1, it's irrelevant.
You didn't followed the Steps provided and given Feedback with the results.
Did you and added them to the Bug Report at least?It's Better to continue on the own's topic and keep relevant info in just one place.
-
Next post are moved to here from https://forum.vivaldi.net/topic/91724/vivaldi-crashes-when-opening-certain-sites-suspecting-sso since that one it's old and already fixed
-
@Zalex108 why are you asking me questions in a thread that is locked? I can't answer there, and I don't see your reply here. I'm completely confused by what is going on.
-
Since you had something to add to the previous topic, it has been reopened.
Your post, moved to the original Topic.
I've linked the solution given by the Original Poster.
-