failover batched dhcpbndupd

[David W. Hankins]

On Fri, Aug 11, 2006 at 07:12:06AM -0400, Nancy Griffeth wrote:
> Doesn't the draft say that it MAY send batched updates?
> Or is it really just that it MAY accept batched update?

I think it just says the format MAY appear this way, either
direction.

ISC DHCP doesn't support this batched bndupd format mechanism,
and the architecture the original author chose (which copies
out values of options present in failover messages into
discrete structures in memory, so there can only be one value
of any option no matter how many times it appears in a failover
message) can't really support it very well.

So I would basically count on ISC DHCP never implementing
this.

Really I think this is a fairly silly thing to put into the
protocol, and is needlessly complicated.  This kind of
behaviour doesn't save you fsync()s, you can just as easily
delay fsync() until read() runs dry.  All it saves is a
tiny number of bytes per binding update (the common/global
options).

And failover isn't bandwidth starved.

-- 
David W. Hankins	"If you don't do it right the first time,
Software Engineer		you'll just have to do it again."
Internet Systems Consortium, Inc.	-- Jack T. Hankins