[bind] Re: Suggestions for coping with this issue

[Tuc at T-B-O-H]

> 
> Tuc at T-B-O-H.NET wrote:
> > [root at ports named]# /usr/sbin/named -v
> > BIND 9.2.4
> > [root at ports named]# rndc flushname www.cnn.com
> > rndc: 'flushname' failed: unknown command
> > 
> > 	Rut roh.... Is that something PAST 9.2.4?
> 
> Yes... 9.4 and above.  As Mark would say:  "upgrade".  :)
> 
> [root at webfarm ~/ISC/bind-9.4.2]# named -v
> BIND 9.4.2
> [root at webfarm ~/ISC/bind-9.4.2]# rndc flushname www.cnn.com
> [root at webfarm ~/ISC/bind-9.4.2]#
>
	I broke the rules and install 9.4.2 ... 

	But it doesn't seem to give the results I want...

	I have the following script run every 3 minutes on 2 machines :

[root at sonar src]# more /usr/local/bin/primecache.sh 
#!/bin/sh
#
# Prime DNS cache for quicker access
#
rndc flushname www.paypal.com 
dig www.paypal.com @localhost > /dev/null
rndc flushname www.paypalobjects.com 
dig www.paypalobjects.com @localhost > /dev/null
[root at sonar src]#  

	But when I dig sonar, and dig ports, I'd expect they'd be
at the same TTL on the records. But they aren't. And when the 3 minutes
are up, it doesn't go back to the full value, just keeps counting down.

	I've run into weird stuff like this before running stuff in
cron. Could it be happening again? I run by hand, runs fine and rests
back constantly. I have the cron as :

#
#
###########################
# Prime the DNS cache     #
###########################
*/3 * * * * /usr/local/bin/primecache.sh > /dev/null 2>&1


I monitor with dig and I get 

241 timeout and 90 seconds to refresh.....
208 and 58 seconds to refresh
180 and 30 seconds to refresh
161 and 10 seconds to refresh
149 after refresh

	Any thoughts whats happening here?

			Thanks, Tuc