Comment 36 for bug 1891657
Revision history for this message
| Brian Murray (brian-murray) wrote Re: [Bug 1891657] Re: systemd 100% cpu usage apport-autoreport.service: Failed with result 'start-limit-hit' | #36 |
On Fri, Sep 25, 2020 at 07:13:52PM -0000, Nicolas Bock wrote:
> On Fri, Sep 25 2020, Brian Murray wrote:
>
> > The issue with PathChanged is that /var/crash is used for writing three
> > types of files. The .crash file, the .upload file (which signifies the
> > crash is ready to be uploaded), and the .uploaded file (which contains
> > the UUID of the crash in the error tracker and indicates the crash has
> > been uploaded). So if PathChanged is used then whoopsie-upload-all will
> > be called for .upload and .uploaded files. Additionally, there is no
> > PathChangedGlob so that we could watch for just .crash files.
>
> Thanks, that makes sense. Although re-reading the
> systemd.path manpage over again I still don't understand how
> this ever worked. It says:
>
> "PathExists= may be used to watch the mere existence of a
> file or directory. If the file specified exists, the
> configured unit is activated."
>
> I read this as saying that the current behavior on Groovy is
> actually correct. Or am I missing something?
A bug was fixed in systemd that resolved a discrepancy between what the
documentation said (which you've quoted) and the actual behavior of
systemd. This is documented upstream in the following issue:
https:/
--
Brian Murray