----- Original Message -----
> QA test 115 is failing in lots of places.
>
> This seems to have been triggered by the systemd changes recently to the init
> scripts, although I suspect some of this may be longer term problems.
>
> Unfortunately the failures are not the same:
>
> vm02 i686 openSUSE 12.1 (Asparagus)
> Starting pmie ... [extra line]
> pmie count after chkconfig pmie off: 1 [not 0]
That first one is a mystery, but..
> vm03 x86_64 Fedora 18 (Spherical Cow)
> vm12 i686 Fedora 17 (Beefy Miracle)
> vm22 x86_64 Fedora 19 (SchrÃdingerâs Cat)
> vm23 i686 Fedora 20 (Heisenbug)
> vm24 i686 openSUSE 13.1 (Bottle)
> # /bin/systemctl enable pmie.service [extra line]
I've seen this one here on one host, and dug deeper recently...
> vm05 i486 Gentoo 2.0.3
> * rc-update: service `pmie' is not in the runlevel `default' [extra line]
> # /sbin/rc-update add pmie [extra line]
> * rc-update: service `pmie' is not in the runlevel `default' [extra line]
>
> I suspect vm02 and vm05 are longer term, and the others are new ... are the
> new ones benign (and should be handled in the filtering) or indicative of
> some problem.
>
> This involves only the pmie init scripts, the others are not showing any
> similar badness.
I noticed pmlogger is missing these checks & verbose helpful reporting (its
only those two that could/should use this diagnostic helper routine). I've
have a separate bug from Marko to look into where we may need to refine the
way we're so sensitive to chkconfig state (and I'm thinking of adding these
more detailed diagnostics into rc_pmlogger to help there) - I believe it's
safe to filter in QA, anyway, if you'd like to do that in the interim.
cheers.
--
Nathan
|