Re: bootstrapping cfengine.

[Daniel Pittman]

On Wed, 12 Nov 2003, Russell Davies wrote:
> I'm currently in the process of deciding how to install cfengine on my
> machines, it seems that simply compiling on the relevant target system
> then stashing the binaries somewhere for redistribution to future
> similar targets, isn't suffcient since the binaries have other package
> dependencies (openssl, berkeleydb).
> 
> This means that I have to solve the problem of general file/package
> distribution before I even start -- but my intentions were to do this
> using cfengine itself, having it look after file distribution to
> remote hosts (of which there may be many). This seems to be an ugly
> circular dependence.
> 
> Obviously this has been solved before, I would like to know how people
> went about this and what is generally regarded as the most elegant
> approach to getting cfengine up and running, static linking?

Deploy a binary with the appropriate dependencies as part of the network
installation of a new OS image to new or replaced hardware. :)

Deal with existing machines by rolling out new managed infrastructure
systems to replace them, with all this in place.

The documentation at <http://www.infrastructures.org/> covers a bunch of
this in more detail.


If you can't do that, treat the initial install of cfengine and it's
dependencies as a one-time operation, and just do it by hand.

Roll an architecture tarball of enough to get the system running, get it
to each machine, and then have that manage the "real" installation.

        Daniel

-- 
One of the chief problems a modern society has to face is how to
provide an outlet for the intellectual's restless energies and yet
deny him power.  How to make and keep him a paper tiger.
        -- Eric Hoffer, quoted in _Eric Hoffer: An American Odyssey_, 1969