Some architectural changes

[Mike McGrath]

On Wed, 8 Oct 2008, Bret McMillan wrote:

> Mike McGrath wrote:
>
> > Well, app1+ are now all identical, for the most part.  This allows us to
> > think of our app servers as truly being part of a farm.  The exception to
> > this right now is those apps requiring our nfs mount in PHX.  I'm still
> > debating what to do about this but so far haven't really come up with a
> > good answer.
>
> 2 choices come to mind:
>
> 1) accept that they're a different category, and partition them accordingly
>
> 2) could maybe tinker around w/ some things like MogileFS or HDFS or something
> even crazier :)
>
> > We'll be distributing load with haproxy.  Just because an app is on a
> > server doesn't mean that it will get hit.  This allows us to distribute
> > load intelligently and hopefully allow us to get local proxies preferring
> > local app servers so geoIP dns can be deployed.
>
> Do all apps get deployed to all servers in this category?
>

Not all apps but most web apps.  In particular this is our 'new fedora
release day' stuff which includes the wiki, docs, get.fp.o, and the mirror
list.  Koji wouldn't be on these hosts nor would asterisk.

> > The bapp servers (there's only 1 right now) will be our job control
> > servers.  Right now we kind of just have cron jobs spread all over the
> > place and in some instances this has caused problems with production
> > traffic.
>
> Cool.  How many bapp servers are we running these days, and how do we choose
> what gets run where?
>

Just one and when bapp2 gets made it'll be more of an 'active' and
'passive' setup.  So all of them will go on one until such a time comes
that we're seeing performance issues then we'll re-evaluate.

	-Mike