netreg Revisited

[Chuck Anderson]

On Fri, Mar 25, 2011 at 10:34:42AM -0400, John Wobus wrote:
> We run a redundant pair and we restart dhcpd every 2
> minutes when there are config changes awaiting
> deployment.  I think we assist ISC by exercising
> dhcpd in a somewhat-extreme manner.

Actually, I think we avoid many dhcpd bugs by restarting so often 
(once every 20 minutes if necessary in our case).  I've read about 
several bugs for long-running dhcpd processes losing failover 
communications, memory leaks, etc. that we have completely avoided...