[Mageia-dev] [RFC] rsyslog vs journalctl

[Colin Guthrie]

'Twas brillig, and Anne Nicolas at 06/02/13 08:43 did gyre and gimble:
> There was a discussion yesterday evening in packager meeting about what
> we should do with rsyslog. It's needed for upgrade from Mageia 2. But
> journalctl is now installed by default.
> 
> Is there some requirement for systemd ? Shall we have both installed? We
> need an answer to deal with upgrade and isos

At present the main systemd package provides "syslog" and also creates
the folder /var/log/journal which is what enables persistent logging
with the journal (without this folder it just logs into /run/log/journal
and keeps it in memory rotating away as needed over time)

I did this to push the journal generally in cauldron and get it used and
see how things go.

If required, I can split this into a separate package
(systemd-persistent-journal-logging) which simply provides "syslog" and
the folder /var/log/journal

This would allow users to pick whether they want persistent logging with
journal or rsyslog (vs. the current choice of only being able to pick
rsyslog *in addition* to persistent logging).


An alternative is to simply document to people how to change the journal
space requirements such that even if persistent logging is enabled, it
will not take up too much room and purge itself quite aggressively.


I suppose the final decision should be based on what kind of experience
we ultimately want to provide in our tools. If I get a chance I can take
a quick look at the drakservices and draklog to see what simple stuff I
could hack in. The options for really nice, service-specific log
searching and viewing are really opened if we force the journal
persistent logs.


So if I were to choose, I'd leave everything as it is. Persistent
journal logging is enabled always but rsyslog is available as an option.

This also has the advantage of meaning we, as bug triagers and
debuggers, always know where to look for extracting the relevant logs.
As Thomas mentioned to me over the weekend, we need a consistent command
to extract the needed info. I said that I would work on ensuring such a
command was included and thus available on the live media too should
people need to use that to boot a broken machine to extract debug info.


But I am also happy to concede to popular opinion here.

Col

-- 

Colin Guthrie
colin(at)mageia.org
http://colin.guthr.ie/

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/