[bind10-dev] Is AXFR/IXFR needed for DB-based (especially large) zones?

[JINMEI Tatuya / 神明達哉]

At Mon, 13 Jun 2011 10:26:34 +0200,
Michal 'vorner' Vaner <michal.vaner at nic.cz> wrote:

> > Looking at other implementations, BIND (9) DLZ insists that it's a bad
> > practice: http://bind-dlz.sourceforge.net/worst_practices.html and
> > database replication should be used instead.
> > 
> > DLZ supports AXFR, too, but if I understand the code correctly, it
> > simply takes the in-memory approach, so it's not scalable for larger
> > zones.  PowerDNS's AXFR support basically seems to be the same (I
> > don't know what's the recommended practice for server synchronization
> > when using PowerDNS).
> 
> Well, the in-memory has a disadvantage that you need to load it manually and it
> takes RAM. We could let the DB backend do something like this for us with stuff
> like copy-on-write versioning. For example, this suggests that mysql is capable
> of doing so:
> http://dev.mysql.com/doc/refman/5.0/en/internal-locking.html
> which I expect happens on disk.

Okay, this seems to be an option.

> Anyway, the 10-minutes lock might not be that big problem maybe, if it was a
> hidden master, then just the write operations would wait for the 10 minutes,
> happen at once and then another sends can happen.

Maybe or maybe not.  Only the users who really manage such a large
zone and want to use a DB backend can provide a useful answer
(developers can only guess, and the result often does not meet real
requirement).  That's why I asked the question here.

---
JINMEI, Tatuya
Internet Systems Consortium, Inc.