PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
19121 darxus 20 0 5666220 1.537g 9468 S 0.0 19.7 105:20.43 Xvfb
23684 darxus 20 0 212908 37372 23372 S 0.0 0.5 0:00.02 Xvfb
The second line is an identically run instance for comparison, run for a minute.
Run as: Xvfb -screen 0 1680x1024x24 -ac -nolisten tcp :5
I'm using this to run cairo's make check once every four hours. No clients currently connected:
darxus@dancer:~$ DISPLAY=:5 xlsclients
For thoroughness, the (only) command I'm running against this display, from cairo's git, is:
DISPLAY=:5 CAIRO_REF_DIR=$HOME/source/cairo-ref.xcb-image.xvfb.script CAIRO_TEST_TARGET=xcb,image,image16 make test
Xvfb has been running for almost five days.
Version 1.18.3, running on Ubuntu 16.04.
I ran Xvfb through valgrind as:
valgrind --leak-check=full --log-file=xvfbvalgrind.log --leak-check=yes Xvfb -screen 0 1680x1024x24 -ac -nolisten tcp :6
Verified that the kill command successfully got it to exit with output.
Ran cairo's make check for about 37 hours (a single run), then ran kill on it, waited three days for it to exit, then had to shut it down so I could move, with no output.
I went back to just using Xvfb normally, problem is still there:
5903 darxus 20 0 6141480 3.046g 9348 S 0.0 39.1 128:20.84 Xvfb
I wanted to mention in an older bug, a workaround was to run Xvfb with -noreset
I haven't tested it.
(Since it took three years for that workaround to get mentioned there.)
-- GitLab Migration Automatic Message --
This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.
You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/xserver/issues/195.