Re: /run/systemd/propagate/example.service deletion

[Lennart Poettering <lennart@xxxxxxxxxxxxxx>]

On Mo, 12.09.22 08:13, Stefan Catargiu (stefan.catargiu@xxxxxx) wrote:

> Hello all,
>
> I have noticed that when using certain sandboxing features for units, e.g. ProtectHome,
> a directory will get created in /run/systemd/propagate with the name of the service,
> e.g. /run/systemd/propagate/example.service, which systemd is then using for certain bind mounts.
>
> Now, the thing is, that directory is never going to be deleted after the service stops,
> which is all good, after all /run is a tmpfs, but this is becoming slightly problematic
> when using instantiated services, you can end up with large numbers of directories
> under /run/systemd/propagate.
>
> I have seen some extreme cases where /run runs out of inodes because of this.
> One extreme example : way too many directories are created under /run/systemd/propagate when a lot
> of coredumps are generated on a system which uses systemd-coredump.
> You will have one instantiated unit per coredump, hence a directory like
> /run/systemd/propagate/systemd-coredump@1-1234-0.service is going to be created and so on.
>
> All things considered, shouldn’t these directories be deleted after a service stops?

THis is probably a bug. Can you please file an issue on systemd github
about this?

https://github.com/systemd/systemd/issues/new?assignees=&labels=bug+%F0%9F%90%9B&template=bug_report.yml

Lennart

--
Lennart Poettering, Berlin