DNS server crashing, what to do?

I have a basic LAMP server, with CPANEL installed, centOS. I am not an admin guy, just a programmer.

But my DNS service keeps crashing, and I need to restart it through cpanel about every 3 days or so.

How do I debug this? Where do I start?

What log files are on centOS? Have you looked there? Are there any other errors?

I don't really know where to start with that. cpanel logs everything, I just don't know where to look...

I'm a programmer, and don't really have much admin knowledge. I guess today is the day I learn...

Bigweb is really good at this sort of thing. I'm sure when he sees this he'll be able to offer some advice.

camperjohn, do you know how to login to your server via ssh?

If you can you need to look at your server log file:

/var/log/messages

You can either view it with pico, vi, less, or more, or type a command like this:

cat /var/log/messages | grep named

So that everything related to DNS will show up. Then see if you notice any errors. Best time to look for the errors is right after this problem happens as they should be towards the bottom of the log file then.

Also if you are able to login please let us know what version of DNS you are using by typing:

/usr/sbin/named -v

/var/log/messages is what I was thinking. I use FreeBSD & didn't know if CentOS was the same.

The messages file mostly consists of the following:

Sep 6 06:45:51 srv1 named[479]: lame server resolving 'sonicspeedsters.com' (in 'sonicspeedsters.com'?): 204.15.192.197#53
Sep 6 06:45:51 srv1 named[479]: lame server resolving 'sonicspeedsters.com' (in 'sonicspeedsters.com'?): 204.15.192.198#53
Sep 6 06:45:51 srv1 named[479]: lame server resolving 'homepagepal.com' (in 'homepagepal.com'?): 204.15.192.197#53
Sep 6 06:45:51 srv1 named[479]: lame server resolving 'homepagepal.com' (in 'homepagepal.com'?): 204.15.192.198#53

It goes forever...

The version I am using is:

root@srv1 [~]# /usr/sbin/named -v
BIND 9.2.4

What next?

Should there be sooo many entries of "lame server resolving..." in the messages file? It is 52 megs in size, just since Sept 6... There are also messages.1, .2, .3, .4 etc, each between 32 and 70 megs.

Update the version of BIND. As I said I don't use CentOS but according to one site, you can try :
yum update bind\*

Yeah your version is a little outdated. Don's advice should work as well to get your BIND updated. If you are using Cpanel I would recommend updating all of your software by typing this in SSH (that would update everything for you that is outdated including BIND):

/scripts/upcp

That will update all of your software automatically for you. Watch the output for any errors on upgrading anything. If there are errors it could be preventing your system from automatically staying updated, and we will need to fix those.

Also to stop the logging of lameservers, which I actually do as well what you need to do is edit:

/etc/named.conf

There should be a section called logging. Mine has this structure:



Inside that logging block I have this line:



Once you add that to that file you can then restart named:

/sbin/service named restart

Let me know if you have any problems following something or need more help editing files using pico or vi