Maintained by: NLnet Labs

[Unbound-users] patch implementing round robin rrsets

Måns Nilsson
Wed Mar 7 22:52:13 CET 2012


On Wed, Mar 07, 2012 at 06:27:08PM +0100, Felix Schueren wrote:
> Hi,
> 
> actually, I feel that DNS-based "load balancing" or "load distribution"
> when relying on RRset randomization is the poorest choice you can make -
> you have no control at all over distribution of the traffic, which means
> that a single system must be able to take the full load anyway, plus you
> end up with downtime of at least RRset TTL, or minimum cache TTL in some
> cases. 

Actually, in our setup, with BIND as resolver we saw close to uniform
load on the two service servers that shared the round-robin record. With
unbound, (still using BIND as name server in the shape of InfoBlox
devices) we had a serious tendency towards loading the machine with the
lowest IP address much harder; more like 80/20 than 50/50. Since this
was the older machine, and we'd relied on round-robin for load sharing,
things went pear shaped PDQ. We threw hardware on the problem, so are good
for now, but round-robin really would set things straight. 

I'm quite happy that the userbase made the effort, and I'm equally delighted
that it seems to be headed into a future unbound release. Open Source win. 

Wouter, any guesstimates on release schedule? 

-- 
Måns, will have something really good to report at morning coffee
	tomorrow.