You are not logged in.
Hi all & HNY!
Since a couple of days after setting up Akonadi on a postgres basis (system-wide instance), starting a KDE session on my desktop computer really sucks because it has become way laggy, there's permanent disk activity, and memory gets eaten up little by little (starting w/ 2G from 8G max [yes, I know my box is old]). Even spawning konsole is occasionally taking ages.
htop is always stating on first position:
/sbin/init resume /dev/mapper/magicbox-swap
But why is there resuming at all now? The machine is no laptop, I don't make any use of hibernation or so (and can't find anything related in KDEs power management). And I am explicitely starting empty KDE sessions.
Admittingly I am quite new to KDE/Plasma, so maybe I missed something in the Plasma configs. – Same goes for PostgreSQL for Akonadi btw., I just thought it's lighter than mariadb. Anyway I tried editing /var/lib/postgres/data/postgresql.conf according to #Troubleshooting might help – but no. Also there's no change when stopping (and then restarting) akonadi.
Further I disabled Qt journal logging behaviour; still no dice.
Then I checked if there's a difference between graphical logging out/rebooting, and doing so in an X terminal – there isn't.
So what's going on here?
Last edited by cameo (2022-02-08 18:46:43)
Offline
/dev/mapper/magicbox-swap is your actual swap partition?
cat /proc/cmdline # do you have the resume stanza in the commandline
ps aux # what runs that process in what context
Ceterum censeo: do you have a parallel windows installation (because that might be hibernating, causing the system to attempt a resume, but on bogus swap data)
Offline
/dev/mapper/magicbox-swap is your actual swap partition?
Yes. But almost zero is in use, while RAM since start up has already eaten up a couple of GBs.
cat /proc/cmdline # do you have the resume stanza in the commandline
In fact I do. And there was never a problem with it. – Shall I remove it anyway?
ps aux # what runs that process in what context
Still only this one:
root 1 0.0 0.1 166772 12252 ? Ss 14:13 0:01 /sbin/init resume /dev/mapper/workbox-swap
Ceterum censeo: do you have a parallel windows installation (because that might be hibernating, causing the system to attempt a resume, but on bogus swap data)
No, I don't, Arch is the only OS.
Meanwhile I don't think this issue is related to KDE (directly).
This heavy 'resuming' even starts before logging in. After some minutes it stops, then continues in X, and won't stop; I fear it will wear out my disk I just bought a year ago.
Even it is related to Postgres – or to btrfs, which has been my file system of choice for years.
This Arch install itself is from July btw., and I am updating it almost daily. – Maybe something went wrong?
Last edited by cameo (2022-01-19 14:50:48)
Offline
Hibernate the system, resume, reboot.
Situation gone?
If not, remove the resume stanza, reboot - maybe that's not the issue at all and the init process is busy w/ something entirely different.
Offline
Well I don't now how to resume this machine acutally; it's a desktop, and I've never this this on purpose.
However, I remember there was a grub update lately.
So I am now going to reboot, edit the command line, delete "resume", and continue ...
Offline
"systemctl hibernate", when you power on the machine it should™ resume the hibernated system.
Offline
OK, to finish this one:
It had nothing to do with resuming.
Further "htop" was the wrong monitoring tool, too. – Spawning "top" instead then revealed a freaking out "baloo" being the culprit!
Disabling it made things calm down again, and that's still the status quo since a couple of weeks. ^^
Offline