Weird state after package and kernel update on Manjaro

Hi there,

this morning I ran updated some Linux packages on my machine. Manjaro also recommended to switch to the latest kernel (from 6.6 to 6.8) which I did since I hardly had issues doing this.

After that (it’s a correlation, not necessarily a reason) I noticed that some of my favorite apps did not auto-start. And I was not able to start them manually.

From the journald protocol I got the following:

05.04.24 13:26	systemd	Started snap.joplin-desktop.joplin-desktop-c683e7c7-9302-4c6c-b7ad-c3ffa06a684f.scope.
05.04.24 13:26	kernel	audit: type=1400 audit(1712316375.255:159): apparmor="DENIED" operation="capable" class="cap" profile="/usr/lib/snapd/snap-confine" pid=30712 comm="snap-confine" capability=12  capname="net_admin"
05.04.24 13:26	kernel	audit: type=1400 audit(1712316375.255:160): apparmor="DENIED" operation="capable" class="cap" profile="/usr/lib/snapd/snap-confine" pid=30712 comm="snap-confine" capability=38  capname="perfmon"
05.04.24 13:26	kernel	audit: type=1400 audit(1712316375.259:161): apparmor="DENIED" operation="open" class="file" profile="snap-update-ns.joplin-desktop" name="/proc/30733/maps" pid=30733 comm="5" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0
05.04.24 13:26	plasmashell	error: system does not fully support snapd: cannot mount squashfs image using "squashfs":
05.04.24 13:26	plasmashell	-----
05.04.24 13:26	plasmashell	mount: /tmp/syscheck-mountpoint-283568545: /dev/loop43 ist bereits eingehängt oder wird gerade benutzt.
05.04.24 13:26	plasmashell	       dmesg(1) könnte nach einem fehlgeschlagenen mount-Systemaufruf
05.04.24 13:26	plasmashell	       weitere Informationen liefern.
05.04.24 13:26	plasmashell	-----
05.04.24 13:26	plasmashell	ERROR: not connected to the gnome-42-2204 content interface.
05.04.24 13:26	systemd	app-joplin\x2ddesktop_joplin\x2ddesktop-4aa7218fe2ff4562a8125532038adcda.scope: Couldn't move process 30712 to requested cgroup '/user.slice/user-1000.slice/user@1000.service/app.slice/app-joplin\x2ddesktop_joplin\x2ddesktop-4aa7218fe2ff4562a8125532038adcda.scope': No such process
05.04.24 13:26	systemd	app-joplin\x2ddesktop_joplin\x2ddesktop-4aa7218fe2ff4562a8125532038adcda.scope: Failed to add PIDs to scope's control group: No such process
05.04.24 13:26	systemd	app-joplin\x2ddesktop_joplin\x2ddesktop-4aa7218fe2ff4562a8125532038adcda.scope: Failed with result 'resources'.
05.04.24 13:26	systemd	Failed to start Joplin.

So obviously it’s not possible for snap to mount a loop device (for whatever reason, I don’t have a deep understanding of the inner workings). After some DuckDuckGoing I figured out to list the available loop devices, and I can confirm that /dev/loop43 already is in use:

losetup -a | grep loop43                                                                                                                                                                                                                                       
/dev/loop43: []: (/var/lib/snapd/snaps/skype_340.snap)

Since I don’t use Skype anymore I gave removing Skype a try, but this comes with the very same issue:

snap remove skype
error: system does not fully support snapd: cannot mount squashfs image using "squashfs": -----
       mount: /tmp/syscheck-mountpoint-283568545: /dev/loop43 ist bereits eingehängt oder wird
       gerade benutzt.

       dmesg(1) könnte nach einem fehlgeschlagenen mount-Systemaufruf
       weitere Informationen liefern.

       -----

So, something is broken with snap.

On the other hand, some other snaps actually work. I’m able to run and use intellij-idea-ultimate as well as android-studio and drawio.

I have the impression that snap somehow has an invalid loop device counter. A loop device /dev/loop44 does not exist on my system so it’s available. Instead of mounting /dev/loop43 for any command it should just try to mount /dev/loop44. So, I’m tempted to manually remove the Skype snap to allow snap to use /dev/loop43 again.

But I did not do so since my above assumption might be pretty stupid. And I don’t want to break the snap setup by manually removing some directories I additionally need to guess.

Would be great if somebody could get me on the right track. I would love to see snap working again as usual.

This sounds like an issue someone reported to LP: Bug #2058848 “Snap does not work with kernel 6.8” : Bugs : snapd which is apparently fixed with 6.8.3 kernel.

1 Like

Thanks for that hint. I was not able to find this one.

Unfortunately, according to the bug it’s occurring again in kernel 6.8.4, which I’m running.