https://bugs.kde.org/show_bug.cgi?id=306352
--- Comment #26 from Timo Gurr <timo.g...@gmail.com> ---
While certainly not a fix by any means I want to share a workaround because I
found myself ever so often not noticing that the cpu spike again kicked in
until I wondered why my machine is loud again and revving up the fans. I
installed monit and dropped a configuration file in /etc/monit.d/ which
monitors the process kglobalaccel5 and restarts it under my user session when
two consecutive checks find the kglobalaccel5 process running with higher cpu
usage than usual/normal:
/etc/monit.d/kglobalaccel5
check process kglobalaccel5
matching "kglobalaccel5"
start program = "/usr/bin/systemctl --machine tgurr@.host --user start
plasma-kglobalaccel.service"
stop program = "/usr/bin/systemctl --machine tgurr@.host --user stop
plasma-kglobalaccel.service"
restart program = "/usr/host/bin/systemctl --machine tgurr@.host --user
restart plasma-kglobalaccel.service"
if cpu usage > 10% for 2 cycles then restart
You may have to adjust the systemctl path and the user of course and it only
works for one user this way. For me this means I don't have to manually care
about monitoring the process anymore and not having to issue the restart
command after every single login all the time.
journal entries should look something along the line like:
Sep 25 20:25:05 exherbo monit[2412]: 'kglobalaccel5' process is running with
pid 2457
Sep 25 20:25:05 exherbo monit[2412]: 'kglobalaccel5' cpu usage of 32.7% matches
resource limit [cpu usage > 10.0%]
Sep 25 20:25:34 exherbo monit[2412]: 'kglobalaccel5' cpu usage of 33.9% matches
resource limit [cpu usage > 10.0%]
Sep 25 20:25:34 exherbo monit[2412]: 'kglobalaccel5' trying to restart
Sep 25 20:25:34 exherbo monit[2412]: 'kglobalaccel5' restart:
'/usr/bin/systemctl --machine tgurr@.host --user restart
plasma-kglobalaccel.service'
Hope it may be useful for someone else as well.
--
You are receiving this mail because:
You are watching all bug changes.
