refresh times out from Win DNS

[Mark Jeftovic]

We've been seeing this sporadically for awhile and it is possible this
happens more often with Win DNS masters (we're not sure but the latest
case is definitely a WIN DNS box)

We are able to do the first transfer OK after which point subsequent
refreshes fail with the usual complaint of

Jul 12 07:23:26 ds2 named[1879]: zone example.com/IN: refresh:
failure  trying master 10.2.229.181#53: timed out

Left to its own the zone eventually expires.

The thing is, we can do AXFR and IXFR from the command line just fine
using host or dig. Also, the slave is not clogged up with transfers
in progress (there are 6 SOA queries in progress and 0 xfers running
as I type this, on a slave with approx. 85K zones configured).

This is bind9.2.3

Any ideas?

-mark

-- 
mark jeftovic