[Mageia-dev] Broken database problem solved but not all well

Colin Guthrie mageia at colin.guthr.ie
Mon Apr 16 15:41:27 CEST 2012

'Twas brillig, and Thierry Vignaud at 16/04/12 14:34 did gyre and gimble:
> BTW why not displaying a message "switch to tty2 for a login"?

In the normal case of X or prefdm.service failing, this is exactly what
happens (it's actually a message saying "your graphical login is broken
blah blah, press any key to switch to multi-user.target (aka runlevel 3".

This  case is different (if I'm correct in my assumption). The fact is
that the non-lsb legacy initscripts impose a very, very strict ordering
on startup (based on their priority numbers in /etc/rcX.d/S??foo filenames).

This strict order is typically full of loops (a -> b -> c -> a), so
somthing has to give. In order to do that systemd will delete jobs to
break the loop. Frequently this job turns out to be "the prefdm.server
start job". When this happens we simply don't know that things are
failing in order ot display the error message (likewise, if we added a
totally separate unit for this it too could be deleted by these jobs).

Ultimately the correctly solution is to provide native systemd or
LSB-compliant sysvinit scripts.

I've fixed a couple and have previously linked a spreadsheet where
everyone can help out and track progress etc.



Colin Guthrie

Day Job:
  Tribalogic Limited http://www.tribalogic.net/
Open Source:
  Mageia Contributor http://www.mageia.org/
  PulseAudio Hacker http://www.pulseaudio.org/
  Trac Hacker http://trac.edgewall.org/

More information about the Mageia-dev mailing list