IXFR and AXFR

[Mark Andrews]

In message <e39f60e8c9db5f989a369841cd92a6e1.squirrel at webmail.aminor.no>, "Eivind Ols
en" writes:
> > At what time the slave executes AXFR and at what time it executes IXFR
> > from the master?
> 
> Someone please correct me if I give misleading information. I don't
> believe I am, but I've been wrong before :D
> 
> There's a good section about this in the ARM, such as BIND 9.7 ARM section
> 4.3 - Incremental Zone Transfers (IXFR).
> 
> Basically, a BIND 9 slave will normally ask for IXFR unless told not to
> (request-ixfr).
> A BIND 9 master can't always provide IXFR though - if it can't it will
> provide AXFR instead. To be able to provide IXFR it needs to have some
> idea of the changes being made, so it can give a meaningful reply when
> asked to provide "all changes since serial number X", so you'll normally
> see IXFR being possible for dynamically updated zones (and a couple of
> other cases, check the ARM).
> 
> Regards
> Eivind Olsen

named will do a axfr initially, anytime it believes it has lost sync with the
master (the ixfr did not apply without error), when "rndc retransfer" is
called, when ixfr is rejected by the master.

The master will return a AXFR style IXFR whenever it doesn't have the requested
axfr stream.

> _______________________________________________
> bind-users mailing list
> bind-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/bind-users
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org