faster propagation

Kevin Darcy kcd at daimlerchrysler.com
Fri Apr 5 01:45:33 UTC 2002


Are you familiar with the NOTIFY protocol? When a zone is reloaded or changed,
a NOTIFY packet is sent to all of the slaves in the NS records, which triggers
them to do a refresh shortly thereafter. Even if you have slaves which are not
in the NS records of the zone -- so-called "stealth" slaves -- you can arrange
for them to receive NOTIFY packets using the "also-notify" configuration
option. NOTIFY is the best way to speed up master/slave replication. I

That having been said, I can't think of any reason other than network problems
and/or machine capacity problems, that replication would take longer than the
REFRESH setting. Is there anything unusual in the master's or slaves' logs to
indicate problems?


- Kevin

ben wrote:

> Is there any way to speed up propagation of a change in a domain record?
> Will changing the REFRESH parameter in the zone file do anything?  My
> refresh is 21600 (6 hours), but it just seems seems to take so long (over 24
> hours), when at my old company, it happened in about 2 hours.  I'm running
> BIND 9 on RedHat 7.1.
>
> Are there even any little tweaks that can be done to speed up the
> propagation?
> Also, is there a better or faster way to find out if/where the change has
> propagated? (besides nslookup and dig?).
>
> ben.



More information about the bind-users mailing list