NFS

[Edward Glowacki]

On 26 Jan 2001, Ben Pfaff wrote:
> That's my assumption too.  But in theory and on average, won't
> this scheme work just as well whether the machines occasionally
> change servers or not?  Of course, if you get pathological access
> patterns, it could be a problem.
> 
> (If you did get them to occasionally re-look-up the IP address,
> you could also hack BIND to, e.g., return the IP address of the
> machine that currently has the lower load average.)

Apparently this is all academic, or moot, now.  If we do push these
boxen into production as NFS servers, we'll probably be upgrading
them to a gigabit NIC, at which point the problem goes away...

-- 
Edward Glowacki			glowack2@msu.edu
Technical Support Services		
Michigan State University