[systemd-devel] Confused about failed service reporting

[Edward Z. Yang]

OK, in that case it sounds like a bug in yum-updatesd, in that it is
dying without reporting an appropriate error code.  This is vexing,
because when I look at the systemd logs I see no entries from yum-updatesd,
and when I start it up after boot, it runs fine (implying that the
ordering dependencies are wrong.)

(Actually, I misdiagnosed our problem, and it's not due to the bug
I linked. As far as we can tell it is an open bug.)

Cheers,
Edward

Excerpts from Lennart Poettering's message of Mon Mar 26 12:03:45 -0400 2012:
> On Mon, 26.03.12 03:33, Edward Z. Yang (ezyang at MIT.EDU) wrote:
> 
> > A few weeks ago, yum-updatesd died due to the following bug
> > https://bugzilla.redhat.com/show_bug.cgi?id=709315
> > 
> > This is exactly something that we would like systemd to report
> > us.  However, while interacting with servers with the dead
> > yum-updatesd, we found it didn't show up when we ran
> > systemctl --failed.  When we ran normal diagnostics, we got:
> > 
> > [root at better-mousetrap ~]# systemctl status yum-updatesd.service
> > yum-updatesd.service - YUM Package  Update Service
> >       Loaded: loaded (/lib/systemd/system/yum-updatesd.service)
> >       Active: inactive (dead)
> >       CGroup: name=systemd:/system/yum-updatesd.service
> > 
> > What's up with that? (Is this yet another "upgrade systemd,
> > and all ye problems will go away?")
> 
> If I read the bug report properly yum-updatesd simply didn't detect
> properly whether a network connection was around. 
> 
> systemd will only store the execution results of services, and if they
> don't actually report any errors then we won't record that as
> "failed". A daemon may report an error by crashing, by timing out or by
> returning with a non-zero exit code.
> 
> Lennart
>