NSlookup query - odd results

I'm running two slave name servers, one running Bind 9.2.4 on Solaris 8 (SPARC, single processor), the other Bind 9.3.2 on Solaris 10 (Sparc, dual Processor). I run nslookup from a Windows XP client in debug mode, looking up the name www.cnn.com. The 9.3.2 server _sometimes_ times out with a "DNS Request Timed out, 2 seconds", but the 9.2.4 server never times out.

So I look at the results a bit closer and noticed a difference. With the
9.2.4 server, the 'authoritative name server' records TTL always refreshes when the A records TTL does. Notice in my output the A record TTL reaching
1 second and the NS records TTL at 8m21s. Next, the A record TTL is refreshed to 5 minutes as well as the NS record TTL back to 10 minutes.

The same query to the 9.3.2 server yields different results. When the A record TTL reaches zero and refreshes, the NS record TTL continues to count down to zero instead of refreshing. When it reaches zero is when I get the timeout on the next query. The proceeding query usual succeeds as do the rest.

Can anyone explain the difference?

  Latest articles

Intranet hostname lookup only

memory consumption limit

special characters in zone files

Slave bind skips delegation record in master zone

Slave zones not updating

SPF RRType

Trying to get full domain info in nslookup