Internet DNS forwarders causing Intranet DNS servers to be unresponsive?

[paulette mudrey]

We have 2 intranet DNS servers that point to our 2 internet DNS servers as 
forwarders and we have noticed that if we have problems with our Internet 
connection that the intranet DNS servers are unresponsive.  In one case we 
got the following errors and had to restart named:

Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 172.17.100.9#53: 
no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 
172.17.133.5#10001: no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 
10.0.227.214#1829: no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 10.0.165.55#1737: 
no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 
172.16.1.31#36332: no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 172.31.101.16#53: 
no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 
172.17.133.5#10001: no more recursive clients: out of memory
Apr 21 14:12:15 hpadm1 /usr/local/sbin/named[1033]: client 10.1.87.157#3486: 
no more recursive clients: out of memory
Apr 21 14:12:14 hpadm1 /usr/local/sbin/named[1033]: client 172.17.100.9#53: 
no more recursive clients: out of memory
Apr 21 14:12:15 hpadm1 /usr/local/sbin/named[1033]: client 10.1.86.149#1031: 
no more recursive clients: out of memory
Apr 21 14:12:15 hpadm1 /usr/local/sbin/named[1033]: client 10.0.168.30#2617: 
no more recursive clients: out of memory
Apr 21 14:12:15 hpadm1 /usr/local/sbin/named[1033]: client 10.0.75.37#1030: 
no more recursive clients: out of memory
Apr 21 14:12:15 hpadm1 /usr/local/sbin/named[1033]: client 172.31.101.16#53: 
no more recursive clients: out of memory
Apr 21 14:12:15 hpadm1 /usr/local/sbin/named[1033]: client 10.1.87.96#4206: 
no more recursive clients: out of memory
Apr 21 14:39:43 hpadm1 /usr/local/sbin/named[1033]: mem.c:1540: 
INSIST(mpctx->allocated > 0) failed

In another case the intranet DNS servers would not even reply to internal 
DNS queries or for anything in it's cache.  I read in the "DNS and BIND"  
book (fourth edition - p. 334 Single point of failure) that this should not 
be the case and that they should reply to anything they have in their cache 
or that they are authoritative for.  Could they just be too busy trying to 
reply to internet DNS queries that are taking longer than usual?  Is there a 
better way to set this up??

We are running BIND 9.1.3 on HPUX 11.0.

Paulette

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE! 
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/