[rancid] Re: post-install, now what?

[Ed Ravin]

On Thu, May 25, 2006 at 11:45:09AM -0400, jim bartus wrote:
>    I've gotten a trial install of rancid going on my network now checking
>    about 20 devices (all cisco, switches, routers, and two pixes).  My
>    question is... now what?
>    For instance, in my old setup (pancho/snmp/tftp based) I had a copy of
>    running config sitting in a tftp root, which made it easy to "copy
>    tftp run" from a device to restore a config.  How do you guys address
>    this in rancid?

We don't.  RANCID post-processes the config in various ways that improve
change reporting but lose the original config.  For starters, by default
all passwords get removed from the config so that they don't accidently
get emailed out, but there are more subtle transformations: sequence numbers
get removed, some things get sorted, etc.  You probably want to keep Pancho
around so that you'll have an exact copy of the config to restore from.

>    I notice that since what rancid stores in cvs is one
>    giant file, I'd have to write a script to parse out the parts I want.
>    Is there a community site of some sort where scripts like these are
>    posted?

The current copy of the config is kept available for your use  - look
in <routergroup>/configs .  You may also want to read through the Perl
code for the  *rancid scripts to see how they parse out bits and pieces
of the config.

>    Also, has anyone done any integration work with syslog so that rancid
>    will automatically run against a device when a log message about its
>    config changing comes in?

See the entry for Simple Event Coordinator (SEC) in the RANCID FAQ:

   http://www.shrubbery.net/rancid/FAQ

>    Mostly I'm just trying to avoid as much re-inventing of the wheel as
>    possible, so I'm hoping to find documentation, tips, best practices,
>    and scripts that may already exist.

The best resource seems to be the archives of this mailing list, where I
got the tip above about SEC :-).