Re: load balancing again

3 msg	problem with views
4 msg	RCODE (REFUSED)
11 msg	Reverse zone with subnet larger than Class C
3 msg	A record Question

load balancing again
\ Pavel Urban (27 Feb 2006)
. \ bdyslm (27 Feb 2006)
. \ Mark Andrews (27 Feb 2006)
. \ Wiley Sanders (28 Feb 2006)
. . \ Kevin Darcy (28 Feb 2006)
. . . \ Merton Campbell Crockett (28 Feb 2006)
. . . . \ Joseph S D Yao (28 Feb 2006)
. \ nick.seccombe (2 Mar 2006)

2 msg	can't resolve similar addresses
8 msg	More than I can bite off...
2 msg	lwres interaction with round-robin
7 msg	dns delegation and recursion
1 msg	Kerenel Parameter Tuning
3 msg	threads problem with 9.3.2
2 msg	Load Balancing
2 msg	Question about forwarder
5 msg	Filtering RR on slave server.
2 msg	BIND Suitability For Providing Failover Web Ser...
8 msg	Bind 9 slaves - new zones are never picked up
3 msg	Interpreting query logs
2 msg	still running 9.2.3...what should I upgrade to?
2 msg	Lock specific records in dynamic zone...
2 msg	reverse lookup example

Subject:	Re: load balancing again
Group:	Bind-users
From:	Joseph S D Yao
Date:	28 Feb 2006

On Mon, Feb 27, 2006 at 07:31:09PM -0800, Merton Campbell Crockett wrote:
...
> Aren't "traffic spikes" just a "fact of life"?
>
> BIND can only exert some control over the sequence in which the name
> servers are listed when a system requests NS records directly from
> your server. Once the list has been delivered to another name server
> or a client system, the decision of how the list will be used is
> based on that name server's or client system's strategy for using the
> list of name servers.
...

Traffic spikes are not "just a fact of life", if one has true load
balancing using a system designed to get feedback from the servers and
allocate new traffic accordingly. At least, this smooths out the
spikes.

But this system is not DNS or BIND.

--
Joe Yao
-----------------------------------------------------------------------
This message is not an official statement of OSIS Center policies.