Re: [Cfengine-develop] Development plan / meeting

[Hugo Gayosso]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark Burgess <address@hidden> writes:

> WHAT??????  Why on earth would you consider using tripwire for
> anything?  It doesn't do anything that cfengine doesn't already do
> "better".

For one, I started using tripwire before I started using cfengine, so
some of my machines were already running it.

Also, one of the features I don't know how to do with cfengine is that
it tells me when/what new files have been added to a directory or
which ones have been deleted.  I could probably hack something with
scripts but tripwire already does that.

Now, maybe cfengine can also do that, but we go again to the point
that a lot of people have made, we need a "real life" examples posted
possibly with comments.  I saw that there is a "tripwire
functionality" posted in the 'hints' page but it does not address what
I described.

Besides, I used 'tripwire' just as an example, another big one is
software management, I settled for using AutoRPM and have cfengine
manage its configuration, plus some directory structure and symlinks.

AutoRPM does its job fine, but a more coupled integration with
cfengine would be nicer, on the other hand I don't think that we
should have cfengine brew coffee and make toast, but a more flexible
plug-in architecture could be useful so maybe somebody will write
something that fulfills a need.

I participated in the Netsaint project (now Nagios) and that was one
of the things that I loved about it, it had a few "core" programs
which had all the "intelligence" and control, but all the service
checks (it is a systems/services monitoring system) were made out of
plug-ins, there were some guidelines that the plug-in developers
needed to follow, and as long as you followed them, your plug-in could
be written in any language, do any kind of check, etc, and Netsaint
will act properly.

Netsaint also has an "event handler" feature, which means that if the
plug-in reports a "WARNING" or a "CRITICAL" Netsaint executes a
user-defined command to try to fix that.  So if I already have this
why I did bother learning cfengine?

Well, because of the autonomous part, Netsaint relies in a central
server and I don't like that, I prefer that the machines take care of
themselves individually or at least as a group, specially because I am
thinking in deploying some GNU/Linux-based laptops which could be
unconnected from the network for extended periods of time.

Also, I don't want to saturate the network with "I am OK" messages
from all machines to a central server, I prefer to be notified of "I
am NOT OK" messages, or better yet, the machines finds a problem and
fixes it, and I don't even know (well, maybe if I look in the logs).


- -- 
Hugo Gayosso
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.0 (GNU/Linux)

iD8DBQE+ZBxyMNObVRBZveYRArmFAJ0Zg3QCSAH0HQ6QDlD9y6TjTLA2UACghcrZ
6Wx3etEh4g2yo9n0k89R1EY=
=4pR/
-----END PGP SIGNATURE-----