> 2. How does mirroring/load balancing work in Debian? Anyone used GNU
> Queue? or mod_backhand?
You didn't mention what services the machines are going to be
providing... you might choose different clustering technologies
depending on what you're doing with the machines.
For example, say you're just doing web serving of mostly static
content... I'd probably just go with a round-robin DNS setup because
it's easy and works fairly well. If one of your boxes fails, you
simply add an IP alias to the interface on the good machine with the
dead machine's address and nothing else in your infrastructure has to
change. It gets more complicated if you need to do a lot of writes on
a web site, have a database back end that needs to update transactions
on the other host(s) in the cluster, etc.
If the machines are going to be file servers, you might choose a
different mechanism. Mail servers, compute servers, etc. There are
solutions, but to guide you, we need to know more about your
environment.
BTW, I'm not sure what things are like with BYU Purchasing lately (it's
been quite a few years since I worked there), but you might find it
harder to buy stuff that's not coming from an "approved vendor".... I
remember the hoops we had to go through to buy SGI workstations.
Purchasing: "Get three bids from three different suppliers."
Me: "The equipment is only available directly from SGI."
Purchasing: "Get three bids from three different suppliers."
... repeat ...
I did finally hook up with a buyer who understood my problem and was
able to help me, but it took a physical visit to the purchasing
department to get it resolved.
-jan-
--
Jan L. Peterson
Unemployed "Computer Facilitator"
http://www.peterson.ath.cx/~jlp/resume.html
____________________
BYU Unix Users Group
http://uug.byu.edu/
___________________________________________________________________
List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
