[rancid] Policy rollout [repost]
paul at gear.dyndns.org
Fri Apr 8 07:11:17 UTC 2011
(It appears this didn't go through when i first posted it - apologies if
you get two copies.)
I've recently discovered and installed RANCID to run my mostly HP
ProCurve network. I'm hoping to replace HP's slow, unreliable, and
overpriced ProCurve Manager Plus with the combination of Observium,
RANCID, and possibly sFlowTrend (or other tools better suited to the
task - suggestions gratefully accepted).
One of the things that is hinted at in some of the howtos is the
suggestion that configurations can be rolled out from the RANCID server.
However, all the documentation i could find about this was the
suggestion to stick the commands in a file and run
hlogin -x filename switchname
Has anyone considered anything a little more sophisticated? I'm
thinking a small extension to the .cloginrc file format would make
sense, where we could have
add policy switchregex groupname path/relativeto/groupdir
Then have rancid-run (or the underlying infrastructure commands - i
haven't traced through them yet) automatically run those command files
before it collects the configuration.
Any pointers on where to start hacking (or where to look to save
reinventing the wheel) would be appreciated.
More information about the Rancid-discuss