Re: [systemd-devel] Use of SystemKeepFree=

2018-10-11 Thread Lennart Poettering
On Do, 11.10.18 12:34, Jan Synacek (jsyna...@redhat.com) wrote:

> Hello all,
> 
> looking at the current code, SystemKeepFree= is not accounted for when
> doing vacuuming, only SystemMaxUse= is used. There was an ancient
> RHEL-7 bug for systemd-219 with the exact same problem. Now I'm not
> sure if that's actually a problem or not, but the documentation
> suggests that SystemKeepFree= should be honored.
> 
> Is it a bug?
> 
> When is SystemKeepFree= actually used?
> 
> Why have SystemKeepFree= at all if it's the "other way around" of
> SystemMaxUse= ?

The vacuum logic should respect SystemKeepFree= correctly afaics.

cache_space_refresh() calculates a single "limit" value from the
available disk space and disk usage, his is then used in do_vacuum()
and passed to journal_directory_vacuum(). Everything appears in order
to me?

This code hasn't been touched in quite some time, hence I'd assume
this always was the way it was?

Lennart

-- 
Lennart Poettering, Red Hat
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/systemd-devel


[systemd-devel] Use of SystemKeepFree=

2018-10-11 Thread Jan Synacek
Hello all,

looking at the current code, SystemKeepFree= is not accounted for when
doing vacuuming, only SystemMaxUse= is used. There was an ancient
RHEL-7 bug for systemd-219 with the exact same problem. Now I'm not
sure if that's actually a problem or not, but the documentation
suggests that SystemKeepFree= should be honored.

Is it a bug?

When is SystemKeepFree= actually used?

Why have SystemKeepFree= at all if it's the "other way around" of
SystemMaxUse= ?

Cheers,
-- 
Jan Synacek
Software Engineer, Red Hat
___
systemd-devel mailing list
systemd-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/systemd-devel