At 3:32 PM -0400 7/4/00, Andrew C. Greenberg wrote:
>Yes, I've given this some thought.  We should be able to quickly 
>build an easily extensible network monitoring and dashboard package, 
>complete with e-mail reporting and beeper capabilities, akin to, 
>say, NetSaint (an EXCELLENT package by the way).

I wonder if STOMP (http://coweb.cc.gatech.edu/csl/27) the testing 
suite that Brian Trammell assembled might not already have all the 
main pieces -- it can already check servers and return errors.

>
>In this way, whomever is responsible for maintaining the system will 
>know immediately, rather than waiting for the first complaints to 
>dribble in from the peanut gallery.

Unfortunately, we don't have anyone "hired" (even at the College 
level) to provide 24-7 response on down systems.  It's just too 
expensive.  But there is *almost* always somebody in the 
Collaborative Software Lab (where our servers live), so a note to 
swiki-bugs usually gets systems back up in a few hours.

>
>Next step: can we build a redundant slave server that can mirror the 
>Swiki and take over the IP address if a heartbeat from the primary 
>server fails -- now THAT would be something.  Indeed, perhaps we can 
>arrange for it to do reasonable load-balancing in the interim?

That can be trickier for us, since our IP's are controlled by a level 
above us, but it's certainly possible.

Mark

P.S. Je77 beat me in this morning, and already re-booted Minnow.

--------------------------
Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280
Associate Professor - Learning Sciences & Technologies.
Collaborative Software Lab - http://coweb.cc.gatech.edu/csl/
(404) 894-5618 : Fax (404) 894-0673 : guzdial at cc.gatech.edu
http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html