Description of problem: After some update my i5 used only 2 out of 4 cores. Even "make -j5" did not push the other 2 cores, however there was some minor load on those cores. Version-Release number of selected component (if applicable): How reproducible: Always. Steps to Reproduce: 1. Add "CPUAffinity=1 2" to /etc/systemd/system.conf 2. Reboot 3. run e.g. "make -j5" in kernel source directory 4. use htop to see only cores 1 and 2 to be under load The problem was that this was quite difficult to debug. I started to search for cgroups and cpuset since that was the obvious choice. No luck there, and the help and tools for debugging cgroups or cpuset problems are virtually non existent. So, the problem was easy to fix when I managed to find the source. Fix: comment out the line saying "CPUAffinity=1 2" in /etc/systemd/system.conf This report is just to leave a record, as requested in Mageia forum, on the behaviour if someone else runs into this kind of problem.
CC: (none) => wilcal.int
Blocks: (none) => 2120
(In reply to comment #0) > > This report is just to leave a record, as requested in Mageia forum, on the > behaviour if someone else runs into this kind of problem. Thanks, Jari :) Assigning anyway, to make certain the maintainers of systemd will know about this issue
Assignee: bugsquad => dmorganecCC: (none) => mageia, marja11
Hmm, the man page suggests this is for the init process specifically. Can I ask what Desktop Environment you are using? Perhaps you are using e.g. KDE which does not properly detach it's processes from the init context. Perhaps GNOME behaves better in this case?
I'm using KDE but the problem was that during some update (or user invention) the "CPUAffinity=1 2" was added to /etc/systemd/system.conf file. The reason for this bug is that someone asked to report it here for future reference.
what about this bug with systemd 40 ?
This was not bug report on mailfunctioning system rather a note for anyone who has same kind of problem so it can be closed.
(In reply to comment #5) > This was not bug report on mailfunctioning system rather a note for anyone who > has same kind of problem so it can be closed Yes, we know, but please answer anyway. If you still run cauldron: is the problem still there in fully updated cauldron? If so, please give the output of: rpm -qa | grep systemd
Keywords: (none) => NEEDINFO
No, the problem is not there anymore.
(In reply to comment #7) > No, the problem is not there anymore. Thx, so closing
Status: NEW => RESOLVEDResolution: (none) => FIXED