https://bugs.kde.org/show_bug.cgi?id=403563
Raman Gupta <rocketra...@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rocketra...@gmail.com
--- Comment #16 from Raman Gupta <rocketra...@gmail.com> ---
This is the issue that comes up when Googling "plasmashell memory leak", so
I'll post here. There are usually multiple root causes of these types of
problems, but without better ways to debug / investigate these issues than "try
it with this configuration" it is really hard for users to determine whether we
are talking about the same issue or not.
On my Fedora 36 system with 128 GB of real memory, systemd-oomd has killed
plasmashell twice in two days for using almost 133 GB of memory (real memory +
considerable swap). There is no conceivable situation in which plasmashell
should be using that much memory.
Oct 23 03:04:40 edison systemd-oomd[1888]: Killed
/user.slice/user-1000.slice/user@1000.service/session.slice/plasma-plasmashell.service
due to memory used (132956889088) / total (134229807104) and swap used
(27105050624) / total (30064758784) being more than 90.00%
Oct 23 03:04:40 edison systemd[2255]: plasma-plasmashell.service: systemd-oomd
killed 199 process(es) in this unit.
and then almost exactly 24 hours later:
Oct 24 03:05:46 edison systemd-oomd[1888]: Killed
/user.slice/user-1000.slice/user@1000.service/session.slice/plasma-plasmashell.service
due to memory used (132685852672) / total (134229807104) and swap used
(27061297152) / total (30064758784) being more than 90.00%
Oct 24 03:05:46 edison systemd[2255]: plasma-plasmashell.service: systemd-oomd
killed 156 process(es) in this unit.
I do find it suspicious that both instances happened at almost exactly the same
time overnight -- the cron log does not show any process triggering at or
around that time.
If there is a way to effectively debug this sort of thing, that would be
helpful.
I don't use slideshow wallpaper.
Output of qdbus supportInformation: https://invent.kde.org/-/snippets/2387
--
You are receiving this mail because:
You are watching all bug changes.
