Puppet users - Feb 2012 - Best sequence for building/managing servers with Puppet?

I need to both build some new servers and rebuild some existing ones
keeping in mind I want to use Puppet. The question is which is the
best sequence to proceed with?

A) Develop modules one by one and make sure each component works, then
assemble and test modules that need to talk to each other and make
sure those work together, all *before* building the servers. Then once
all the modules are completed and tested for each piece of
configuration and functionality needed at the systems level they can
be used to build each server automatically.

B) Build the servers manually first and retrofit them with pre-
existing Puppet modules, hoping that the modules will just work AND
match how each system should be for your particular needs.

C) Build the servers manually first and then develop the modules
needed to manage each one.

Is there one way that saves more time? Is there one that lends a
higher degree of confidence everything will work as expected at the
end?

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Feb 7, 2012, at 1:46 PM, weloki wrote:
> I need to both build some new servers and rebuild some existing ones
> keeping in mind I want to use Puppet. The question is which is the
> best sequence to proceed with?
> 
> A) Develop modules one by one and make sure each component works, then
> assemble and test modules that need to talk to each other and make
> sure those work together, all *before* building the servers. Then once
> all the modules are completed and tested for each piece of
> configuration and functionality needed at the systems level they can
> be used to build each server automatically.
> 
> B) Build the servers manually first and retrofit them with pre-
> existing Puppet modules, hoping that the modules will just work AND
> match how each system should be for your particular needs.
> 
> C) Build the servers manually first and then develop the modules
> needed to manage each one.
> 
> Is there one way that saves more time? Is there one that lends a
> higher degree of confidence everything will work as expected at the
> end?
----
I don''t think there''s an easy answer and it probably relates
more to the way you operate. At some point you will get a core set of modules
and then you can put it into action.

I found it easiest to use VMPlayer on my desktop (Macintosh) and setup the
puppet master (always running) and spin up new VM installs time and again until
I got everything fairly well tuned and my need to refactor diminished to almost
none. I think the refactoring thing is key. When you start with puppet, you will
want to refactor pretty much everything for a while as your understanding
becomes deeper. Once you have deployed modules, refactoring becomes a much more
difficult task. But it will become obvious when you are ready to start
deploying.

Craig

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Feb 7, 2:46 pm, weloki <wel...@gmail.com>
wrote:
> I need to both build some new servers and rebuild some existing ones
> keeping in mind I want to use Puppet. The question is which is the
> best sequence to proceed with?
>
> A) Develop modules one by one and make sure each component works, then
> assemble and test modules that need to talk to each other and make
> sure those work together, all *before* building the servers. Then once
> all the modules are completed and tested for each piece of
> configuration and functionality needed at the systems level they can
> be used to build each server automatically.
>
> B) Build the servers manually first and retrofit them with pre-
> existing Puppet modules, hoping that the modules will just work AND
> match how each system should be for your particular needs.
>
> C) Build the servers manually first and then develop the modules
> needed to manage each one.
>
> Is there one way that saves more time? Is there one that lends a
> higher degree of confidence everything will work as expected at the
> end?

If you don''t yet have any Puppet infrastructure then the shortest time
to achieving operational servers is provided by (C), perhaps with a
touch of (B) mixed in.  Really, that''s the only approach that makes
sense unless you have a development lab in which to build and test
your new Puppet infrastructure before deploying it, and enough time to
use it.  A development lab for this purpose could be virtual (e.g.
Craig''s VMs), but you''re still looking at non-trivial time to
get an
interesting configuration together.

As I already hinted, I don''t think (B) and (C) are exclusive.  If
you''re open to using third-party modules at all, then you should
definitely seek existing modules that fulfill some of your
requirements.  As things currently stand, however, you should be
prepared to tweak such modules for your own needs and for
compatibility with other modules (whether local ones or other third-
party ones).


John

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.