I am running the latest beta on both server and client, on 64 bit ubuntu 1.0-20160913r13691-1. Both server and client are running gnome-flashback desktop with metacity window manager.
I experience server hang and crash. The symptom is that sometimes while connecting to the server, suddenly all forwarded windows disappears and the connection is closed. Subsequent connection to the server is not possible. On server xpra list show server running but cannot be contacted. xpra stop :100 will stop the server, but X.org and sound server will be left running and have to be killed separately.
On the other hand, sometimes after quitting the server and reconnecting, the windows become unresponsive. I cannot click or open anything and server have to be killed manually, as above.
I did xpra -d all on both server and client on one of these ocassions, and I append the result to this ticket
server log result of xpra -d all
client log with xpra -d all
The first two attached files are from an instance of unable to reconnect after disconnecting. The second pairs of attached logs are from actually abruptly quitting while connected to xpra server and then server becoming unresponsive.
I also observe that these crashes only appear in beta versions starting this September. The version I used throughout August, xpra_1.0-20160806r13238-1, is rock solid. However, it does not forward notification properly, even after I set firefox to use gnome native notification, per bug 1304.
Now I am forced to revert to the August beta, because I'm running a production system and cannot afford the constant crashes... notification is a small issue by comparison.
xpra stop :100 will stop the server, but X.org and sound server will be left running and have to be killed separately.
FYI you should be able to:
There are quite a few changesets between r13691 and r13238... and I'm not seeing the crash. I've even tried running gtkperf in a tight loop for hours (that's a sort of torture test for xpra) Can you try the r13570 build so I know if the bug was introduced before or after that? What do I need to do to reproduce the bug? This is an X11 crash, usually caused by code that talks to X11 from the wrong thread. The two main classes show a lot of changes (mostly for scrolling support), but nothing stands out: window_source.py diff, window_video_source.py diff
Thanks for the tip! I'm running r13691 on client now and the notification is back! On server, I tried to run r13570, but the server failed to start. (error at the end of the log).
Right now, I'm running r13671 on the server to see if it crashes... I will report back after a day or two if it does, and sooner if it does.
BTW, the window fix you introduced on r13671 does make a big difference in making firefox extension forecastfox (fixed) working, as well as all sorts of context/popup window in firefox works much better (it doesn't get truncated or split between two screens anymore) compared with, say, r13238. Only by reverting back to the old way did I see how terrible was the original behavior! Thank you!
XPRA_CLIPBOARD_LIMIT=30 xpra --xvfb='Xorg -noreset -nolisten tcp +extension GLX +extension RANDR +extension RENDER -logfile ${HOME}/.xpra/Xvfb-10.log -config ${HOME}/.xpra/xorg.conf' start :100 --bind-tcp=0.0.0.0:10000 Warning: Ubuntu 'trusty' breaks Xorg/Xdummy usage - using Xvfb fallback [Errno 2] No such file or directory: '/home/qian2/~/.xpra' xpra main error: Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/xpra/scripts/main.py", line 124, in main return run_mode(script_file, err, options, args, mode, defaults) File "/usr/lib/python2.7/dist-packages/xpra/scripts/main.py", line 1132, in run_mode return run_server(error_cb, options, mode, script_file, args, current_display) File "/usr/lib/python2.7/dist-packages/xpra/scripts/server.py", line 1004, in run_server logfd = open_log_file(log_filename0) File "/usr/lib/python2.7/dist-packages/xpra/scripts/server.py", line 553, in open_log_file return os.open(logpath, os.O_WRONLY | os.O_CREAT | os.O_TRUNC, 0o666) OSError: [Errno 2] No such file or directory: '/home/qian2/~/.xpra/:100.log'
qian2@workstation:$ XPRA_CLIPBOARD_LIMIT=30 xpra --xvfb='Xorg -noreset -nolisten tcp +extension GLX +extension RANDR +extension RENDER -logfile ${HOME}/.xpra/Xvfb-10.log -config ${HOME}/.xpra/xorg.conf' start :100 --bind-tcp=0.0.0.0:10000 Warning: Ubuntu 'trusty' breaks Xorg/Xdummy usage - using Xvfb fallback [Errno 2] No such file or directory: '/home/qian2/~/.xpra' xpra main error: Traceback (most recent call last): File "/usr/lib/python2.7/dist-packages/xpra/scripts/main.py", line 124, in main return run_mode(script_file, err, options, args, mode, defaults) File "/usr/lib/python2.7/dist-packages/xpra/scripts/main.py", line 1132, in run_mode return run_server(error_cb, options, mode, script_file, args, current_display) File "/usr/lib/python2.7/dist-packages/xpra/scripts/server.py", line 1004, in run_server logfd = open_log_file(log_filename0) File "/usr/lib/python2.7/dist-packages/xpra/scripts/server.py", line 553, in open_log_file return os.open(logpath, os.O_WRONLY | os.O_CREAT | os.O_TRUNC, 0o666) OSError: [Errno 2] No such file or directory: '/home/qian2/~/.xpra/:100.log'
Thanks to @afarr for testing #770 and reminding me of another important change: r13600 changed some focus code, and one of the places where this is called is on disconnection (which fits with your description from comment:1). r13769 fixes that. I believe this particular bug is now fixed, I have posted some updated trusty beta builds. Let me know if you manage to get any sort of crashes with the 1.0 branch.
PS: the startup bug (introduced in r13507) you reported in comment:3 had already been fixed in r13593.
I am fairly confident that this is fixed so closing it for now, let's re-open if someone hits this again.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/1307