Call for Help: Setting up Saltstack for devops of cluster

Craig Rodrigues rodrigc at
Mon Dec 1 21:13:56 UTC 2014

On Sun, Nov 30, 2014 at 9:39 PM, Jason Wolfe <jason at> wrote:

> Craig,
> Did we have one of havoc/wreck/chaos earmarked to function as the master,
> or did we want an isolated instance running the orchestrating? The load for
> a handful of machines should be next to nothing, so dedicated hardware is
> not required, but an isolated environment would likely be best for various
> reasons including security.
> Getting the framework set up should be a day or so of work, but dialing in
> all of the logic to replicate the current state of things and the requests
> in the gist will be a bit longer.  With Salt Helium having just gone gold,
> I think we have a solid release to get things moving.
> Jason

We don't have a host earmarked as the master.  For now can we pick "havoc"?
Is it relatively easy to change later on?


On Mon, Dec 1, 2014 at 10:52 AM, Johannes Meixner <xmj at> wrote:

> Do we have a description on which host functions how? I.e. which of those
> mentioned is salt-master, which are buildslaves?
> Are there already any salt files? Do we want to use the Jenkins formula
>, how much do we
> need to
> extend it, or sould we start writing our own NIH solution ? ;-)


We have a list of machines in the Jenkins cluster here:
We don't have a designated saltstack master, but for now I would say we can
use "havoc".

In terms of the saltstack Jenkins formula, I don't know if it makes since
to use it or not.

Right now, the main Jenkins host ( ) has Jenkins via:
   -> pkg install jenkins
           -> some entries in rc.conf like jenkins_enable="YES" is also running an nginx server configured with a
reverse proxy:

Take a look at
carefully and look at those requirements.


More information about the freebsd-testing mailing list