[systemd-devel] [ANNOUNCE] systemd 27

[Ran Benita]

On Sat, May 21, 2011 at 15:53, microcai <microcai at fedoraproject.org> wrote:
>
> 于 2011年05月21日 20:44, Kay Sievers 写道:
> >>> No, NO, NO! Don't use the release.
> >>>
> >>> It's totally unstable. do some systemctl <TAB> stuff, systemd crashed.
> >>
> >> Hmm?
> >>
> >> In which way?
> >>
> >> Can you provide a backtrace?
> >
> > systemd 27 works fine here, the tab-completion too. Seems more like a
> > local problem.
>
> Try more times. say, 100 times. systemd will crash eventually.

If systemd crashes, I believe the kernel should panic; is that what happened?

On another note, here's something nice to try (as a regular user):

echo -e does-not-exist{1..131072}.service\\n | xargs -P4 systemctl load

(see MANAGER_MAX_NAMES in manager.h).
You won't be able to do much with systemctl afterwards (Failed to
issue method call: Argument list too long), besides reboot --force.

Am I missing something? Is there a way to trigger a garbage collection
on the useless units?