Slave DNS server hangs during restart (Bind)

[mangia]

Hi

I have two DNS servers in master - slave configuration

When I restart master, everything is OK but slave server sometimes hangs.

Logs:

Mar  7 01:55:27 ns2 named[1712]: shutting down: flushing changes
Mar  7 01:55:27 ns2 named[1712]: stopping command channel on 127.0.0.1#953
Mar  7 01:55:27 ns2 named[1712]: stopping command channel on ::1#953
Mar  7 01:55:27 ns2 named[1712]: no longer listening on 127.0.0.1#53
Mar  7 01:55:27 ns2 named[1712]: no longer listening on 192.168.100.9#53

After that, bind doesn't work and I can see that it is still running and it does not respond to queries....

[root@ns2 ~]# ps ax|grep named
 1712 ?        Ssl    0:02 /usr/sbin/named -u named -t /var/named/chroot
 2488 pts/0    R+     0:00 grep named
[root@ns2 ~]#

Logs from master server are similar but after I stop bind I have a line

Mar  6 23:58:39 ns1 named[10781]: exiting

As you can see above, this line doesn't exist in slave server log

Commands pkill named, service named restart, service named stop,... doesn't help..... Only option is to force server restart... this is not a solution because server needs 5-10 min. to boot up again...

Centos, BIND is chrooted, v 9.3.4

Any ideas ?

[falko]

What's the output of

netstat -tap

after you've tried to restart BIND? I guess there's still a BIND process running.

[mangia]

Hi... Sorry for digging out but the problem is still here

output of netstat -tap doesn't show named (or bind) as running... All I can see is http, mysql and few other services ...

[falko]

Can you post the full output?

[mangia]

I can't post output because I receive the next error

// ***************
1. You have included 12 images in your message. You are limited to using 4 images so please go back and correct the problem and then continue again.

Images include use of smilies, the BB code [img] tag and HTML <img> tags. The use of these is all subject to them being enabled by the administrator.

/ ****************

Of course, I didn't post any images

The only difference when BIND works and in the moment it doesn't work is the next few lines visible in the moment after reboot when everything works fine

tcp 0 0 localhost6.localdomain:rndc *:* LISTEN 2129/named
tcp 0 0 localhost.localdomain:rndc *:* LISTEN 2129/named
tcp 0 0 localhost.localdomai:domain *:* LISTEN 2129/named
tcp 0 0 172.16.162.8:domain *:* LISTEN 2129/named
tcp 0 0 ns2.DOMAIN.COM:domain *:* LISTEN 2129/named

[falko]

Originally Posted by mangia

I can't post output because I receive the next error

// ***************
1. You have included 12 images in your message. You are limited to using 4 images so please go back and correct the problem and then continue again.

Images include use of smilies, the BB code [img] tag and HTML <img> tags. The use of these is all subject to them being enabled by the administrator.

/ ****************

Of course, I didn't post any images

You can use an SSH client such as PuTTY and copy&paste the output.

[mangia]

Hi

If you look my first post, you'll see that I'm able to use putty ...

Just like I said above, posting putty output in new post was followed by the error that I have to many images inside my post what is ridiculous.

Here is the difference when named works

tcp 0 0 localhost6.localdomain:rndc *:* LISTEN 2129/named
tcp 0 0 localhost.localdomain:rndc *:* LISTEN 2129/named
tcp 0 0 localhost.localdomai:domain *:* LISTEN 2129/named
tcp 0 0 172.16.162.8:domain *:* LISTEN 2129/named
tcp 0 0 ns2.DOMAIN.COM:domain *:* LISTEN 2129/named

and when it doesn't work, I don't see the lines posted above...

[falko]

Does

netstat -tap | grep domain

give you any output when BIND isn't working?

[mangia]

Tried that but "domain" or "named" does not exist any more in list...

[falko]

And there are no errors in your logs?