On Sep 28, 2010, at 8:41, James Turnbull <[email protected]> wrote:

> Brice Figureau wrote:
>>
>> I think my reluctancy comes from adding a new dependency to the whole
>> chain (both the stomp client library on every clients _and_ something
>> like ActiveMQ somewhere). As long we still support HTTP with no extra
>> dependency then I think we're fine.
>
> This is always one of my concerns.  Actually what'd it be great to align
> to (albeit I am not sure how feasible it is) would be:
>
> 1.  Scrap webrick as base server and replace with xxxx -
> mongrel/passenger/other?

Webrick sucks but it's the only option that doesn't induce more
dependencies, right?

> 2.  Have async queued mode as the scaling mode

I might rather say 'production' than scaling, but that's my basic
goal.  However, I don't think itMs quite that simple; at least, we've
still got some complexity around adding async to the indirector, and
only when we've kind of sorted that can the story become clear.

> This would mean rather than people going from webrick (which sucks
> obviously), to passenger/mongrel and then potentially to a queuing model
> we'd only have two modes - normal and "wow you've got a lot of
> nodes/configuration".



-- 
Luke Kanies | +1-615-594-8199

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

Reply via email to