Problem with 9.1.1

[Mark.Andrews at nominum.com]

> 
> Hi, i have 3 ns, 1 primary, 2 slaves, running Bind 9.1.1 on solaris 7.
> This is the problem.
> 
> I get this two errors on log:
> 
> Apr 18 00:24:17.298 general: warning: dns_master_load: slave/macrosa.com.ar:1
> 0: $TTL 3306160128 > MAXTTL, setting $TTL to 0
> Apr 18 00:24:26.515 general: info: running

	Well give it a sane TTL value.  Anything above 7 days is self
	defeating as most nameservers use 7 days as a upper bound on
	the TTL when it is cached.
> 
> After that, all the zones ~2000, got a refresh_talkback, this happens everyti
> me i start the named.
> The thing is, that the 3 ns are in the same vlan, with no firewall or anythin
> g blocking the communication on port 53,
> so i dont know why it says time out for 200.45.0.114 (the primary) (the slave
> s have other primarys for some zones, these zones also got this error)
> 
> Apr 18 00:24:45.090 general: info: refresh_callback: zone energia.com.ar/IN: 
> failure for 200.45.0.114#53: timed out
> Apr 18 00:24:45.610 general: info: refresh_callback: zone grupotradis.com.ar/
> IN: failure for 200.45.0.114#53: timed out
> Apr 18 00:24:45.610 general: info: refresh_callback: zone travelpass.com.ar/I
> N: failure for 200.45.0.114#53: timed out
> Apr 18 00:24:45.610 general: info: refresh_callback: zone agf-allianz.com.ar/
> IN: failure for 200.45.0.114#53: timed out
> 
> But, when i change a zone, the serial, and hup named on primary works ok repl
> icating the zone changes, so, what is this problem?

	The nameserver is extremely busy at startup and the answers havn't
	arrived / been processed in time.  The nameserver will retry later.
	These messages are only a problem if they continue which from your
	comments they don't.

> 
> This week, a lot of times happened that the service stops responding querys a
> nd i have to do a kill -9 and start it again cause wont respond with a HUP i 
> dont know if this is related to the large 
> volume of errors (refresh_talkback) sended to the log file and this somehow s
> lowing the service performance, and taking it to a still.
> 
> Any help on this problems, specially any idea why the named hangs if its rela
> ted to this, or not, would be great!
> Thanks!

	It sounds like it is running out of per-process memory.  Use
	datasize to up the amount available.

	Mark
> 
>  Federico Bockel
--
Mark Andrews, Nominum Inc.
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark.Andrews at nominum.com