'MAC affinity' doing the exact opposite, leading to 'pool churn'?

[David W. Hankins]

Jesse Norell wins the 'I read the manpage and all I got was this
stupid award' award.  Those config examples were in the manpage. :)

On Mon, Sep 24, 2007 at 11:31:07AM -0600, Jesse Norell wrote:
> for you, mostly out of curiosity.  I don't know why it works, as my
> understanding is that any hba setting just selects what server would
> answer a given client, so one should work as well as another; I just ran

That's correct: all of the examples you cited split the bitspace of
the HBA evenly between primary and secondary.  That is half the bits
are one, half are zero.  The difference is in which precise bits are
set or not.  '55' and 'aa' are binary 1s and 0s following each other
(1010, 0101).  ff:00 is all-ones followed by all-zeroes, obviously.

So these are all different ways of acheiving the same theoretical
thing, which may not be the same actual thing operationally (as is
so often the case).


The thing changing the split or hba would fix is situations that look
like this:

> > 17:35:43 s2 dhcpd: balancing pool 80de6f0 134.58.217/24  total 201  free 117  backup 84  lts -16  max-own (+/-)20
> > 17:35:43 s2 dhcpd: balanced pool 80de6f0 134.58.217/24  total 201  free 121  backup 80  lts -20  max-misbal 30

Where the 'lts' is holding at the 'max-own' limit.  In order to keep
the 'lts' at the limit, the servers have to shift leases that HBA
to themselves...working against affinity.

The thing to do here is either to adjust the hba (preferrable) or
to change (it's configurable) the lease ownership limit.

-- 
Ash bugud-gul durbatuluk agh burzum-ishi krimpatul.
Why settle for the lesser evil?	 https://secure.isc.org/store/t-shirt/
-- 
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