[rancid] Adding more commands during Rancid / Development

[Alan McKinnon]

It's an iterative process where you tweak your subs and regular
expressions until you get the output you want.

Running rancid-run repeatedly puts strain on your routers though, so use
a short cut:

jrancid -d    gives the jlogin command to use
use that and save the output to a file
use the file as input to jrancid -f

Now you can run it as many times as you like without hitting the network

As for writing subs,

add the command to @commandtable with the literal command as a key and a
suitably named sub as the value. The subs all follow this pattern:

 # This routine parses "show chassis environment"
sub ShowChassisEnvironment {
    print STDERR "    In ShowChassisEnvironment: $_" if ($debug);
    s/^[a-z]+@//;
    ProcessHistory("","","","# $_");
    while (<INPUT>) {
        tr/\015//d;
        last if (/^$prompt/);
                               <= regex magic goes here
    }
    return(0);
}


For me the hard part was figuring out how ProcessHistory() works -
specifically what the arguments do. They just mark the output off in
section and classify chunks of output that goes together. Change values
around and runt he code to observe what they do. It's hard to express
briefly in English so I'm not gonna try ;-)



On 18/09/2013 02:36, Jason Biel wrote:
> Understood, but even when in a dev environment, what is the best way to
> test changes to file for added commands?  For example, I want to add new
> commands to jrancid, but I am not 100% sure on the sub routine
> configuration for parsing output. I suppose I can make the changes, use
> rancid-run, and just keep tweaking until the output in the configs/
> directory is what I want
> 
> I'm trying to mentally build out a process of adding commands for the
> different devices we use without just shotgunning it.
> 
> 
> On Tue, Sep 17, 2013 at 9:34 AM, Alan McKinnon <alan.mckinnon at gmail.com
> <mailto:alan.mckinnon at gmail.com>> wrote:
> 
>     On 17/09/2013 13:00, Jason Biel wrote:
>     > While I understand the basics of adding additional commands to RANCID,
>     > what is the ideal way to be testing new commands as I am debugging
>     > them?  Are most people running a dev RANCID install for testing?
> 
>     Most definitely :-)
> 
>     In any kind of real-life scenario, your rancid data in configs/
>     naturally evolves into a very critical data source. Almost nothing else
>     out there is capable of documenting real life systems to the extent
>     rancid can. This is great for your ego, but also means you can't fiddle
>     with it so much.
> 
>     I keep 2 dev installs for different purposes and follow the classic
>     dev/test/debug/deploy strategy.
> 
>     There's a second reason why this is a good idea - it is very common for
>     sysadmins to maintain their own little patchset to customize rancid
>     behaviour (mostly to account for quirks between device models) and then
>     repatch new releases if your tweaks didn't make it into the codebase. A
>     dev setup makes this process so much easier.
> 
> 
>     --
>     --
>     Alan McKinnon
>     alan.mckinnon at gmail.com <mailto:alan.mckinnon at gmail.com>
> 
>     _______________________________________________
>     Rancid-discuss mailing list
>     Rancid-discuss at shrubbery.net <mailto:Rancid-discuss at shrubbery.net>
>     http://www.shrubbery.net/mailman/listinfo/rancid-discuss
> 
> 
> 
> 
> -- 
> Jason
> 
> 
> _______________________________________________
> Rancid-discuss mailing list
> Rancid-discuss at shrubbery.net
> http://www.shrubbery.net/mailman/listinfo/rancid-discuss
> 


-- 
Alan McKinnon
alan.mckinnon at gmail.com