Desktop: linux client crashes for no discernable reason

[lemmy04]

Expected behaviour

I would expect the client not to crash

Actual behaviour

the client crashes

Steps to reproduce

1. install nextcloud client on linux
2. configure syncinc
3. ?
4. client crashes

Client configuration

Client version:
2.5.3 built from source

Operating system:
openSUSE 15.0 / Plasma

OS language:
german

Qt version used by client package (Linux only, see also Settings dialog):
Built from Git revision 477614 using Qt 5.13.0, OpenSSL 1.1.0i-fips 14 Aug 2018

Client package (From Nextcloud or distro) (Linux only):
distro package

Installation path of client:
/usr/bin

Logs

https://gist.github.com/lemmy04/2b3409177a7ebe2efed153a6a97ccede

[giancarlogiesa]

i think to have the same problem with the same 2.5.3
Linux distribution: Debian 10
Installed: from ufficial Debian package
language: italian

the CPU go to 100% and will never end the job or just crash

i think that can be a GUI problem because i read on a Debian's forum that an user can run the client from terminal without problems

[lemmy04]

well i'm having the same problems on windows as well, client crashes quietly without any error message - in fact the only way to know that it crashed is to open the windows system tray, and put your mouse pointer over the nextcloud icon - if it disappears the client had crashed some time ago, and you need to restart it. And there is no way of knowing when it crashed and how much syncing is necessary to get "up to scratch".

[bstin]

I also have Debian 10 and have noticed that after upgrading to 2.5.3 client suddenly the tray icon will disappear and syncs no longer happen (although running ps -ax still shows Nextcloud client running...zombie?)

I have reverted to 2.5.2 for this reason.

[lemmy04]

2.5.2 wasn't any better for me...

[djamps]

Had to go back to 2.3.3. Now stable as a rock.

[Ezwen]

I have a Linux Mint 19 user who experiences multiple silent crashes per day of the desktop client 2.5.3 (flatpak version). I asked him to provide some logs, but the logs don't show anything about the crash, and the last four messages are all of the type [OCC::FolderWatcher::changeDetected Detected changes in paths ....

EDIT 1: My user used gdb and could see the last line of the stack trace (he could not get the entire trace for some reason):

Thread 1 "nextcloud" received signal SIGSEGV, Segmentation fault.
0x00007ffff7dfed88 in OCC::PropagatorCompositeJob::slotSubJobFinished(OCC::SyncFileItem::Status) () from /app/lib/libnextcloudsync.so.0

EDIT 2: my user finally obtained a stack trace:

Thread 1 "nextcloud" received signal SIGSEGV, Segmentation fault.
0x00007ffff7dfff98 in OCC::PropagatorCompositeJob::scheduleSelfOrChild() ()
   from /app/lib/libnextcloudsync.so.0
(gdb) backtrace
#0  0x00007ffff7dfff98 in OCC::PropagatorCompositeJob::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#1  0x00007ffff7e00670 in OCC::PropagateDirectory::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#2  0x00007ffff7dfffb4 in OCC::PropagatorCompositeJob::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#3  0x00007ffff7e00670 in OCC::PropagateDirectory::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#4  0x00007ffff7dfffb4 in OCC::PropagatorCompositeJob::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#5  0x00007ffff7e00670 in OCC::PropagateDirectory::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#6  0x00007ffff7dfffb4 in OCC::PropagatorCompositeJob::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#7  0x00007ffff7e00670 in OCC::PropagateDirectory::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#8  0x00007ffff7dfffb4 in OCC::PropagatorCompositeJob::scheduleSelfOrChild() ()
    at /app/lib/libnextcloudsync.so.0
#9  0x00007ffff7e0092e in OCC::OwncloudPropagator::scheduleNextJobImpl() ()
    at /app/lib/libnextcloudsync.so.0
#10 0x00007ffff013a2e6 in  () at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#11 0x00007ffff012edbb in QObject::event(QEvent*) ()
    at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
--Type <RET> for more, q to quit, c to continue without paging--
#12 0x00007ffff0b725d1 in QApplicationPrivate::notify_helper(QObject*, QEvent*)
    () at /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5
#13 0x00007ffff0b79980 in QApplication::notify(QObject*, QEvent*) ()
    at /usr/lib/x86_64-linux-gnu/libQt5Widgets.so.5
#14 0x00007ffff0104529 in QCoreApplication::notifyInternal2(QObject*, QEvent*)
    () at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#15 0x00007ffff0157603 in QTimerInfoList::activateTimers() ()
    at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#16 0x00007ffff0157e8c in  () at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#17 0x00007fffef40f4ae in g_main_context_dispatch ()
    at /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#18 0x00007fffef40f748 in  () at /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#19 0x00007fffef40f7e0 in g_main_context_iteration ()
    at /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
#20 0x00007ffff01581f3 in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#21 0x00007ffff010312b in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#22 0x00007ffff010b112 in QCoreApplication::exec() ()
    at /usr/lib/x86_64-linux-gnu/libQt5Core.so.5
#23 0x000055555560fb88 in main ()
(gdb) 

[DominiqueFuchs]

The fix for this issue was merged yesterday. If you'd like to try this out, you can install the daily build containing this changes - would be great if someone of the affected users here can report back on this 😄 Thanks to @mawasak !

[lemmy04]

I've just built it myself on OBS, lets see how it works.
Would this also fix the similar crashes happening on windows?

[lemmy04]

looking at the change itself, this might even fix the sync conflicts that I keep having with files that change a lot. now i'm kinda excited.

[DominiqueFuchs]

@lemmy04 May be, but a look at the log files would be required to check if the corresponding error originates from this specific circumstance (regarding the other sync issues). Thanks for reporting back here!

[lemmy04]

well, problem with that is that for some reasons I do not get any logfiles - they are always just empty...

[lemmy04]

hm so the latest windows daily build just crashed "out of the blue".
nextcloud-2.7.0.5832-daily-20190920-Release.exe <-- that one.

where would I find logfiles of they exist?

[DominiqueFuchs]

Sorry that you are experiencing issues with the client build - wondering why bc I just built an installer package from master 10 minutes ago, but on windows. Check this page for ways to obtain logs on the client:

https://docs.nextcloud.com/desktop/2.5/troubleshooting.html#log-files

[lemmy04]

i'm not 100% sure yet but so far the linux client has been behaving ok. lets give this some more time before I close it tho.

[lemmy04]

I've been running the 2.6.0 client for a few days now, no more crashes - fix confirmed.

[TBoneStaek]

I've been running the 2.6.0 client for a few days now, no more crashes - fix confirmed.

You mentioned earlier that you were having this issue on Windows but then mentioned that the Linux client was working for a few days with the 2.6.0 version. I'm having the exact same issue on Windows 10 64 with Nextcloud client version 2.6.4. It will run in system tray and the only way to know if it crashes after a few days is to hover over the icon and then it disappears. Then I have to restart it and it will run a day or so longer, then repeat.