Unable to start bind (server not coming online)

I recently changed ip for one of my server machines, and from thence, unable to get it online.

bind is not starting at all even after repetitive restarts.

even restarted the machine number of times, no chance.

CJ.

Status: 
Active

Comments

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 00:35 Pro Licensee

no errors when validate servers is done

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 09:38 Pro Licensee

also, check bind config returned no errors

When you attempt to start BIND from the command line, do you see any errors? Or does it just silently fail?

Also, do you see any errors in /var/log/messages when you try to start it up?

The first thing you might want to check in case of such errors is console output when you try to manually start BIND, and if anything useful gets logged in /var/log. Linux applications tend to be very verbose about things they do and especially about things that go wrong. :)

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 11:16 Pro Licensee

attempt to start bind from CLI using command service named start or restart simply returns "[FAILED]"

Okay... so, how about the log files, do you see anything logged into /var/log/messages when you attempt to start BIND?

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 12:00 Pro Licensee

var log mess attached

It looks like it did start OK.

If you run ps auxwwww | grep named , is the BIND (named) process shown?

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 12:25 Pro Licensee

output: root 20619 0.0 0.0 61160 752 pts/2 S+ 12:24 0:00 grep named

Is there any chance I could login to your server to take a look?

If that's okay, you can either email me root login details at eric@virtualmin.com, or you can enable Remote Support using the Virtualmin Support module.

Thanks!

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 12:52 Pro Licensee

I cannot access it by domain name nor IP from outside. but i will connect my router to the second NIC and send the details to access the server to your email.

please check your email in a minute from now.

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 12:57 Pro Licensee

i have sent login details to your email. Let me know if you were able to log in or not.

Howdy -- I don't seem to be able to see anything at that IP address (including SSH or Webmin). Actually, I don't seem to be able to ping it either :-)

Is the IP address you mentioned in your email definitely online?

bislinks's picture
Submitted by bislinks on Wed, 03/09/2011 - 16:15 Pro Licensee

i have sent another mail with details

It looks like this error is being generated each time BIND starts:

named[19942] general protection rip:2ad46d1a7d72 rsp:424e6f30 error:0

That error suggests the application is crashing with some form of segfault, and is often a sign of some kind of hardware problem.

My recommendation would be to check the various hardware components involved, and in particular, the RAM in your server. There's various RAM testing tools out there, such as "memtest", which is included on several Linux LiveCD's that you could boot from.

It's also possible that if there's a hard drive issue, the "named" binary could be corrupted.

Joe's picture
Submitted by Joe on Thu, 03/10/2011 - 14:30 Pro Licensee

You can check the integrity of the BIND installation using rpm verify:

rpm -V bind9

If it returns nothing, the package integrity is fine. If it returns changed binary files, it might be indicative of trouble.

Joe's picture
Submitted by Joe on Thu, 03/10/2011 - 14:34 Pro Licensee

Oops, that should be:

rpm -V bind

bislinks's picture
Submitted by bislinks on Thu, 03/10/2011 - 22:52 Pro Licensee

it returned nothing.

Which test returning nothing, the rpm command Joe showed? That's good then, as that means the BIND package isn't corrupt.

Another thing you may want to try is the memtest, since errors like what you're seeing can be RAM related.

The CentOS LiveCD has a copy of memtest on it, which can be used for testing your RAM. I'd suggest letting it do a full test of your RAM to make sure there isn't a problem there.