We've finally, finally finished up getting just about everything here
under config management(in our case puppet), the last step being to get
RANCID up and dealing with a few odd ducks in the network equipment. I
also just finished up the initial work on our new deployment/promotion
system, which is a redmine plugin(you can view it on my github here:
https://github.com/cammoraton/redmine_deployment) that uses mcollective to
trigger various actions on remote machines(mainly svn checkouts, which we
use to update heira, puppet and our yum repository) and which can also tie
into Jenkins/Hudson, and wrapped some things around the vmware API and
cobbler to do automated provisioning of test nodes. There's a ton of work
left to do there and a lot of refactoring required but the basic
underpinnings are done.
So now it's time for the next step, start applying some software release
tricks to our config management infrastructure and then, ideally, to tie
this into jenkins. It seems straightforward enough, but the trick will be
to not have to duplicate everything in order to test everything. I have a
nice broad sketch of what I'd like it do, but of course the devil is in
the details here.
Has anyone on list done this? It seems like the natural evolution of
config management so surely someone has.
--Nick
_______________________________________________
bblisa mailing list
[email protected]
http://www.bblisa.org/mailman/listinfo/bblisa
