Failover strangeness

[David W. Hankins]

On Mon, Oct 16, 2006 at 11:50:13PM +1000, Glenn Satchell wrote:
> You could always write a patch for the man page(s) and submit it. David
> Hankins mentioned a little while ago that he can integrate
> documentation updates very easily.

They don't require review process (it's very hard to own people's
machines with manpages), so I can just proofread them and commit.
Just send diffs to dhcp-bugs at isc.org so we don't forget them.  It
would be helpful if the diffs worked against the latest 3.1.x.

Also: I'm very hesitant to write OMAPI documentation myself considering
our long-term goal is to phase out OMAPI in favor of a more entrenched,
operational, open management framework (candidates are things like
dBus or IETF's zeroconf).

> >   I guess the frustrating part of this is that I have to take this 
> >manual action for the failover to work correctly.  That's problematic.

It is impossible to distinguish between communications-interrupted and
partner-down.  I mean that figuratively.

The simple loss of network comms with a failover server is not
sufficient evidence that it is not actively handing out leases.

This knowledge comes only, with complete certainty, from understanding
that a server has entered SHUTDOWN state, or from an operator who can
physically confirm it.

In some environments, operators have found that it is more or less
a good indicator - where failover pairs are directly connected via
crossover ethernet cables, or where networks are reasonably static,
or even where the potential for harm, having two servers operating
in partner-down, is less than the potential harm of losing dhcp
service.

But this has to be a "I know this works, I choose to automate entry
to partner-down" conscious decision.

> I don't have it with me at the moment, but I wrote a simple shell
> script that ran on the primary server. If it couldn't ping the other
> server for 10 minutes or so then it switched the local server to
> partner down mode using the above logic. If you think this would be
> useful I can dig it out next week and then post it, but the logic is
> very simple to implement.

Hopefully at some point we can make this a configurable feature.

-- 
ISC Training!  October 16-20, 2006, in the San Francisco Bay Area,
covering topics from DNS to DDNS & DHCP.  Email training at isc.org.
-- 
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