[systemd-devel] Antw: [EXT] I/O error on "systemctl kill -s HUP rsyslog.service"

[Lennart Poettering]

On Fr, 14.08.20 06:42, Harald Dunkel (harald.dunkel at aixigo.com) wrote:

> On 8/13/20 11:07 AM, Lennart Poettering wrote:
> >
> > No! It's a bug. Not in systemd, but LXC. But generating errors in such
> > a borked setup is *good*, not bad, and certainly nothing to hide.
> >
>
> Surely its not a bug in systemd, but ignoring unreasonable data (maybe with
> a warning, if necessary) has a proud tradition in computing. Not to mention
> that systemd ignores duplicate PIDs in the same context as well, AFAICS.
>
> Ignoring PID == 0 wouldn't be unreasonable, regardless whose bug this is.

Well, we have to agree to disagree. We won't hide errors like this one
in systemd. We

Also, why even? If we'd ignore this specific error nothing for you
would be fixed, because to host systemd instance and the one in the
container will conflict with their cgroup use all the
time. i.e. remove each other's cgroups at inconenient times, will not
be able to detect correctly when a service is stopped, or get notified
when it shuts down, because the cgroups it cares about won't be empty
when they are supposed to be empty.

Lennart

--
Lennart Poettering, Berlin