Hmm, interestingly I note that I happen to have one of these leftover transient units on my machine:
$ systemctl status snap.zoom-client.hook.configure.16523059-a244-423d-a1a3-40a56bd63587.scope
● snap.zoom-client.hook.configure.16523059-a244-423d-a1a3-40a56bd63587.scope
Loaded: loaded (/run/systemd/transient/snap.zoom-client.hook.configure.16523059-a244-423d-a1a3-40a56bd63587.scope; transient)
Transient: yes
Active: active (running) since Tue 2021-07-20 05:34:19 CDT; 2 days ago
Tasks: 0 (limit: 38185)
Memory: 106.7M
CGroup: /system.slice/snap.zoom-client.hook.configure.16523059-a244-423d-a1a3-40a56bd63587.scope
Jul 20 05:34:19 ubuntu systemd[1]: Started snap.zoom-client.hook.configure.16523059-a244-423d-a1a3-40a56bd63587.scope.
While mine is for a hook rather than an app, it's the same invocation from the snapd side of things so it's equivalent. This is on Hirstute with systemd 247.3-3ubuntu3.4.
I had a poke at the underlying cgroup, here are the results if someone wants to take a look, but AFAICT the cgroup exists because systemd created it and hasn't destroyed it, the issue is that systemd should have destroyed it when the processes in it exited.
Hmm, interestingly I note that I happen to have one of these leftover transient units on my machine:
$ systemctl status snap.zoom- client. hook.configure. 16523059- a244-423d- a1a3-40a56bd635 87.scope client. hook.configure. 16523059- a244-423d- a1a3-40a56bd635 87.scope transient/ snap.zoom- client. hook.configure. 16523059- a244-423d- a1a3-40a56bd635 87.scope; transient) slice/snap. zoom-client. hook.configure. 16523059- a244-423d- a1a3-40a56bd635 87.scope
● snap.zoom-
Loaded: loaded (/run/systemd/
Transient: yes
Active: active (running) since Tue 2021-07-20 05:34:19 CDT; 2 days ago
Tasks: 0 (limit: 38185)
Memory: 106.7M
CGroup: /system.
Jul 20 05:34:19 ubuntu systemd[1]: Started snap.zoom- client. hook.configure. 16523059- a244-423d- a1a3-40a56bd635 87.scope.
While mine is for a hook rather than an app, it's the same invocation from the snapd side of things so it's equivalent. This is on Hirstute with systemd 247.3-3ubuntu3.4.
I had a poke at the underlying cgroup, here are the results if someone wants to take a look, but AFAICT the cgroup exists because systemd created it and hasn't destroyed it, the issue is that systemd should have destroyed it when the processes in it exited.
I'm inclined to say that this is a systemd bug and not a snapd bug. The code for how we create the transient unit can be found here: https:/ /github. com/snapcore/ snapd/blob/ master/ sandbox/ cgroup/ tracking. go#L216- L319