named-xfer troubles

[Kevin Darcy]

Atsushi RYOBE wrote:

>Hello all,
>
>I have troubles about named-xfer (AXFR).
>
>When slave server requests zone transfer, following troubles happen.
>
>A. Slave Server Side  ( BIND-8.3.7 on Solaris 2.6)
>
>   1) Many logs about named-xfer are recorded.
>
>      host named-xfer[pid]: print_output: short answer (80, 216), zone a.b.c
>
>   2) Temporary backup file (a.b.c.XXXXXX) becomes large.
>      SOA RR is recorded many times until the timeout. 
>
>B. Master Server Side ( BIND-8.3.7 on Solaris 2.6)
>
>   1) Many logs about query are recorded.
>
>      data time info: XX /192.168.0.1/a.b.c/SOA/IN
>
>
>I captured packets between slave server and master server, found that
>slave server tried SOA RR querys until the timeout.
>
>When I try SOA RR query and zone transfer using dig command, it ends
>correctly. 
>
>If you have good ideas, whould you give me any advice?
>
We used to see similar behavior occasionally, particularly on large 
zones. It went away, of course, when we upgraded all of our slaves to 
BIND 9 (which doesn't even have a separate named-xfer program). Perhaps 
you should consider doing the same.

- Kevin