hi everyone

I have been using bind 9.3.2 on centos 4.2 then 4.3 then 4.4 for some time now. I started seeing the following behaviour:

bind would get so slow that it impacts on all of the services on whether they have to do with dns or not. When I reboot my server I can wait up to 2 hours for around 25 zones to load. Ridiculous i thought. At first I suspected my configs, but this is not the case as i loaded it onto another server and within 30 second the zones were loaded and bind serving dns requests. the differences between the 2 server as for hardware is none. The one server that works i'll call it svrgood is on a test lab in my office running off a 1mb adsl connection and it has tomcat installed and also serves as a samba server. the problematic server i'll call svrbad. appart from the services above that it does not have it runs nagios. now even nagios suffers under the bind problem. svrbhad is also in a data centre with a fast connection to the internet.

i at first suspected nagios, but it was not as i stopped nagios and the same problem still persisted. even afdter stopping nagios for a day.

here is some server stats:
version
Linux version 2.6.9-42.0.3.ELsmp (buildsvn@build-i386) (gcc version 3.4.6 20060404 (Red Hat 3.4.6-3)) #1 SMP Fri Oct 6 06:21:39 CDT 2006

bind ver
BIND 9.2.4

please can anybody help with this problem of mine.

thanks in advance

kobus

Have you checked your firewall config on the servers? (e.g. iptables). What are you doing for port 53? Port 953 for rndc?

Hi FW allows 53 tcp and udp but nothing for 953. Did nopt think that it need 953 for public access.

Hello,

I think it is kernel problem.

http://www.linuxquestions.org/questi...d.php?t=514887

Also I had dedicated server .. with kernel 2.6.9-42.0.3.ELsmp and apache became very slow periodically.