Slave Zone Issue

13 posts / 0 new
Last post
#1 Mon, 01/17/2011 - 15:04
mrwilder

Slave Zone Issue

Hi,

I'm trying to create a slave zone on a secondary ns. I've added the master zone to the master ns, but when I hit "Apply configuration" nothing happens.

The dns servers are clustered and worked in the past.

When I clicked apply configuration, the logs show:
Jan 17 14:00:06 localhost named[22920]: received SIGHUP signal to reload zones
Jan 17 14:00:06 localhost named[22920]: loading configuration from '/etc/named.conf'
Jan 17 14:00:06 localhost named[22920]: max open files (1024) is smaller than max sockets (4096)
Jan 17 14:00:06 localhost named[22920]: using default UDP/IPv4 port range: [1024, 65535]
Jan 17 14:00:06 localhost named[22920]: using default UDP/IPv6 port range: [1024, 65535]
Jan 17 14:00:06 localhost named[22920]: the working directory is not writable
Jan 17 14:00:06 localhost named[22920]: reloading configuration succeeded
we loaded zones locally here
we loaded zones locally here
we loaded zones locally here
Jan 17 14:00:06 localhost named[22920]: reloading zones succeeded

On the secondary, I see nothing in /var/log/mesg. How do I diagnose this?

Thanks!

Mon, 01/17/2011 - 19:28
Locutus

Does the new zone get created on the cluster slave? You might remove and re-activate its cluster slave status to make sure it's still correctly active.

What distros do the two servers use? On my Ubuntu, the BIND messages end up in /var/log/syslog.

Tue, 01/18/2011 - 02:26
mrwilder

The slave zone was not created on the secondary for the new master zone on the primary.

It used to work... so, I'm pretty sure the problem is not the distro, but one's CentOS, and one's Fedora- don't see a "syslog" on either of them.

I DID change the root password on both machines, I suspect I didn't change it in all the correct places, possibly.

Tue, 01/18/2011 - 04:00
Locutus

Did you try as I suggested and removed the cluster slave and re-added it? Does it still not work then? If not, you might check Webmin's own debug log which you can turn on in its configuration, maybe you get a hint there why the DNS cluster slave commands aren't working.

Wed, 01/19/2011 - 22:59
mrwilder

Well, I deleted the cluster slave, then deleted it from the Webmin Servers index. I added it back to the webmin servers index. Adding it back to the webmin servers index, that worked, but I can't add the slave back to the cluster for some reason...it just hangs with "Adding ns1.domain.com" ... and stops.

Any idea?

Thu, 01/20/2011 - 03:56
Locutus

Hmm.. Did you try logging on to the clustered server at Webmin -> Webmin Servers Index and then clicking the big icon for the server? When you do so, you should see the Webmin panel of the "remote" server, tunneled through the "local" Webmin. If that works, adding it as a Cluster DNS Slave should work too... Is BIND running correctly on the cluster slave? What auth method did you set up for the connection?

Fri, 01/21/2011 - 12:11
mrwilder

When I tunnel through from remote servers index icon, it does indeed work. However, if I click the "edit" link under the icon, it often says the connection timed out for "status".

Bind reports that it is running on both machines.

Link type is Webmin.

What might I have missed?

Fri, 01/21/2011 - 12:25
Locutus

How did you configure the remote link? Can you paste a screenshot of the "Edit Server" screen? There should be no timeouts when connecting to it or editing it. It's especially strange if it only "sometimes" times out. :)

Fri, 01/21/2011 - 13:38
mrwilder

I've changed the domain name and port.

Again, clicking the icon DOES connect to the remote server. Unfortunately, perhaps I am wrong and the status is always timed out on this screen... my memory is bad.

Thanks -

Tony

Fri, 01/21/2011 - 18:33
Locutus

Okay, the issue might be that the hostname cannot be resolved in all cases, which usually is a sign that the resolving machine has one (or more) correct and one (or more) incorrect nameservers configured.

Try using the IP address of the machine in question instead of the hostname, and see if that changes anything.

Sat, 01/22/2011 - 00:43
mrwilder

Thanks. I tried it, but unfortunately for me it still timed out, though.

Sat, 01/22/2011 - 04:30
Locutus

Mmh.. Okay, sorry, short of intermittent actual connection problems I'm currently out of ideas here.

Sat, 01/22/2011 - 18:26
mrwilder

Ok, how about this. I manually created the slave zone. I did not add any records, but just added the master server's IP address and clicked "apply zone"

I get the following error:

This server was restarted OK, but the following errors occurred on slaves : ns2.domain.com : HTTP error : HTTP/1.0 302 Moved Temporarily

That's gotta be a clue to somebody... hehe...maybe even me soon.

Anybody recognize the problem? Could it be that I use a non-10000 port for webmin?

Thanks, Tony

Topic locked