RT farming management?

hello everyone,

first, and as it is my first post, let me introduce myself:

I’m working as computer engineer at Strasbourg University (fr) where i’m
in charge of RT, sympa and Sogo. I’m a old perl monger (since '97) and
the author of few cpan modules (https://metacpan.org/author/MARCC), i
also am a former Koha ILS contributor and patched occasionally other oss
softwares as well as CPAN modules and tools (like dh-make-perl).

Strasbourg university needs to provide more than one RTs to its large
community so we plan to use (and maybe write) a farm manager to ease the
process of deploying and updating lot of RT instances.

an instance would be a rt fcgid server (one per instance) started with fine
environment variables (CONFIG, local::lib, …) behind nginx as a
dispatcher.

this tool must be capable to

  • be installed and setup easlisy from CPAN
  • install multiple versions and patches of RT codebase (running cpanm)
  • easily setup, start, stop, and show the status rt instances
  • easily archive an instance
  • easily share extension to many instances

this tool must be capable to

the configuration of every instance could be tested (can access to DB,
local::lib is fully populated, …) and easily backuped.

everything must be done via cli (optionnally, a little web app to
control the whole thing).

I don’t see another simple solution (please don’t tell me about the
configuration management systems) but maybe i’m wrong so any feedback
(or existing code, experiments,…) is welcomed.

If there is none, i’ll start my project by next week.

regards
Marc Chantreux
Université de Strasbourg, Direction Informatique
14 Rue René Descartes,
67084 STRASBOURG CEDEX
:phone:: 03.68.85.57.40
http://unistra.fr
"Don’t believe everything you read on the Internet"
– Abraham Lincoln

everything must be done via cli (optionnally, a little web app to
control the whole thing).

I don’t see another simple solution (please don’t tell me about the
configuration management systems) but maybe i’m wrong so any feedback
(or existing code, experiments,…) is welcomed.

While we have written one-off scripts for these deployments, and I
know other folks have as well, I suspect that you’ll find that most of
the modern work is being done with a configuration management system.
I’ve not used it, but I know folks have worked on

and there may be a chef equivalent.

-kevin