Hello, I am using Fedora Core 3, 32bit SMP-system, xorg via default Fedora RPMs. My problem is: I am using xvfb for a java-application (ProjectX) to demux multimedia files. And I found out that: If a "nice 19"-process, which takes all the idle-cpu capacity, is running on the server, xvfb do not give the java-application any free cpu capacity. Only if there is "idle" cpu capacity, xvfb give it to the java-application. Example: Start a setiathome client on the server (or any other heavy process) with: nice -n 19 ./setiathome Than this process will use the whole free cpu-capacity if it is not needed (lowest nice-level), so "idle" cpu-capacity will be zero and "nice" cpu-capacity to the maximum available. Now try to start any process with also heavy cpu-usage in xvfb. It will not work ... or better: it will work, but without any cpu-capacity, taking hours to finish :( Even if all available cpu-capacity is only used by a nice-process on the lowest nice-level, xvfb do not take any cpu-capacity from this process to give it to the "non nice"-process which was startet in the xvfb. With XFree xvfb on Fedora Core 1, this worked. Another thing: if you stop all "heavy" nice-processes on the server and the application running in xvfb take all needed cpu-capacity, they are not shown in the "top"-view or with correct cpu-usage under "ps -aux". You can only see that the process is taking cpu-capacity over the "user" cpu-capacity in the "top"-view. Maybe this problem is allready fixed in 6.8.2, I don't know. It would be wonderfull if an update would be available for 6.8.1 :) This is the RPM-version I use: [root@baghira ~]# rpm -q xorg-x11 xorg-x11-6.8.1-12.FC3.21 [root@baghira ~]# rpm -q xorg-x11-Xvfb xorg-x11-Xvfb-6.8.1-12.FC3.21 [root@baghira ~]# Sorry for my english. If you have any questions, please contact me. Thank you, bye Alex
Sorry, forgot to say: xvfb is running under a normal user-id, not as root
niceness has nothing to do with X, blame your kernel's scheduler.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.