I have caching-nameserver running on FC6. No matter what I do it will note recognizea zone set as a Slave.
The whole purpose of this machine is to be a secondary DNS.
Any comments?
Mike Dwiggins
_________________________ This email was transferred using an Office free edition of AXIGEN Mail Server.
Mike Dwiggins wrote:
I have caching-nameserver running on FC6. No matter what I do it will note recognizea zone set as a Slave.
The whole purpose of this machine is to be a secondary DNS.
Any comments?
Yes, a caching nameserver is not a secondary server.
Go to /usr/share/doc/bind-*/arm and read the documentation.
Ed,
Read that. Now for a newbie how do I go from a caching-nameserver to regular bind.
The only way I could get yum to install bind was to call for the caching-nameserver.
Mike Dwiggins
Ed Greshko wrote:
Mike Dwiggins wrote:
I have caching-nameserver running on FC6. No matter what I do it will note recognizea zone set as a Slave.
The whole purpose of this machine is to be a secondary DNS.
Any comments?
Yes, a caching nameserver is not a secondary server.
Go to /usr/share/doc/bind-*/arm and read the documentation.
_________________________ This email was transferred using an Office free edition of AXIGEN Mail Server.
On 3/28/07, Mike Dwiggins mike@azdwiggins.com wrote:
Ed,
Read that. Now for a newbie how do I go from a caching-nameserver to regular bind.
The only way I could get yum to install bind was to call for the caching-nameserver.
As I stated in your other thread: `yum install bind`
Mike Dwiggins
Ed Greshko wrote:
Mike Dwiggins wrote:
I have caching-nameserver running on FC6. No matter what I do it will note recognizea zone set as a Slave.
The whole purpose of this machine is to be a secondary DNS.
Any comments?
Yes, a caching nameserver is not a secondary server.
Go to /usr/share/doc/bind-*/arm and read the documentation.
This email was transferred using an Office free edition of AXIGEN Mail Server.
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
Mike Dwiggins wrote:
Read that. Now for a newbie how do I go from a caching-nameserver to regular bind.
The only way I could get yum to install bind was to call for the caching-nameserver.
If you were to look closely you'd find out that the caching-nameserver contains only the following:
/etc/named.caching-nameserver.conf /etc/named.rfc1912.zones /usr/share/doc/caching-nameserver-9.3.2 /usr/share/doc/caching-nameserver-9.3.2/Copyright /usr/share/doc/caching-nameserver-9.3.2/rfc1912.txt /var/named/localdomain.zone /var/named/localhost.zone /var/named/named.broadcast /var/named/named.ca /var/named/named.ip6.local /var/named/named.local /var/named/named.zero
In other words, the caching-nameserver package contains *only* the configuration files used to make bind a "caching nameserver". In still other words, the only difference between a "caching nameserver" and an "authoritative nameserver" is the configuration file.
I suspect bind is really installed.
So, go back to the documentation and read the part about setting up the configuration file. I believe there is already an example in the doc that has a sample of a server acting as a "slave" server.
Ed
On Wed, 2007-03-28 at 16:46 -0700, Mike Dwiggins wrote:
I have caching-nameserver running on FC6. No matter what I do it will note recognizea zone set as a Slave.
The whole purpose of this machine is to be a secondary DNS.
I think you'll have to explain what you've tried. A very brief synopsis of how I've done what you're trying (run a caching name server, also as a slave server for local DNS) follows:
I added two configuration files, and left the remaining ones as they were:
My "/var/named/chroot/etc/named.conf" file:
---------------[begin paste]------------------ ## LAN:
view lan_resolver { match-clients { localhost; }; match-destinations { localhost; }; include "/etc/lan.conf"; include "/etc/rndc.key"; };
include "/etc/named.caching-nameserver.conf";
## end of file ---------------[end paste]------------------
You'd want to customise where my example uses "localhost" to suit the networks you're going to allow to query the server.
And my "/var/named/chroot/etc/lan.conf" file:
---------------[begin paste]------------------ ## LAN:
zone "lan.example.com" { type slave; file "slaves/lan.example.com.zone"; masters { 192.168.1.2; }; };
zone "1.168.192.in-addr.arpa" { type slave; file "slaves/1.168.192.in-addr.arpa.zone"; masters { 192.168.1.2; }; };
---------------[end paste]------------------
Then my slave zone files get created, automatically in "/var/named/chroot/var/named/slaves". And that's it...
At 2:25 PM +0930 3/29/07, Tim wrote:
On Wed, 2007-03-28 at 16:46 -0700, Mike Dwiggins wrote:
I have caching-nameserver running on FC6. No matter what I do it will note recognizea zone set as a Slave.
The whole purpose of this machine is to be a secondary DNS.
I think you'll have to explain what you've tried. A very brief synopsis of how I've done what you're trying (run a caching name server, also as a slave server for local DNS) follows:
I added two configuration files, and left the remaining ones as they were:
My "/var/named/chroot/etc/named.conf" file:
---------------[begin paste]------------------ ## LAN:
view lan_resolver { match-clients { localhost; }; match-destinations { localhost; }; include "/etc/lan.conf"; include "/etc/rndc.key"; };
include "/etc/named.caching-nameserver.conf";
## end of file ---------------[end paste]------------------
You'd want to customise where my example uses "localhost" to suit the networks you're going to allow to query the server.
And my "/var/named/chroot/etc/lan.conf" file:
---------------[begin paste]------------------ ## LAN:
zone "lan.example.com" { type slave; file "slaves/lan.example.com.zone"; masters { 192.168.1.2; }; };
zone "1.168.192.in-addr.arpa" { type slave; file "slaves/1.168.192.in-addr.arpa.zone"; masters { 192.168.1.2; }; };
---------------[end paste]------------------
Then my slave zone files get created, automatically in "/var/named/chroot/var/named/slaves". And that's it...
Tim -- is this also a solution for my problem? I have set up a local server (on CentOS 4.4) to test a replacement for a real server with a domain name (running RH 7.2). The local server should always resolve that domain name to itself, so as to properly test itself and not the real server. Just adding lines to /etc/hosts was insufficient, probably because DNS queries (dig, nslookup) still see the real server. I'm trying to avoid really understanding BIND at this time, as I have plenty of other issues to deal with, and the new server, when deployed, won't need this testing DNS configuration.
Tony Nelson wrote:
Tim -- is this also a solution for my problem? I have set up a local server (on CentOS 4.4) to test a replacement for a real server with a domain name (running RH 7.2). The local server should always resolve that domain name to itself, so as to properly test itself and not the real server. Just adding lines to /etc/hosts was insufficient, probably because DNS queries (dig, nslookup) still see the real server. I'm trying to avoid really understanding BIND at this time, as I have plenty of other issues to deal with, and the new server, when deployed, won't need this testing DNS configuration.
If you don't want to understand named and how to configure it then try the following.
1. Remove the /etc/resolv.conf 2. Create a /etc/hosts file that contains all of the information needed and that includes all of the servers that host may need to contact for testing purposes. 3. Edit the /etc/nsswitch and change the "hosts" line to have "files" only.
At 6:31 AM +0800 3/30/07, Ed Greshko wrote:
Tony Nelson wrote:
Tim -- is this also a solution for my problem? I have set up a local server (on CentOS 4.4) to test a replacement for a real server with a domain name (running RH 7.2). The local server should always resolve that domain name to itself, so as to properly test itself and not the real server. Just adding lines to /etc/hosts was insufficient, probably because DNS queries (dig, nslookup) still see the real server. I'm trying to avoid really understanding BIND at this time, as I have plenty of other issues to deal with, and the new server, when deployed, won't need this testing DNS configuration.
If you don't want to understand named and how to configure it then try the following.
- Remove the /etc/resolv.conf
- Create a /etc/hosts file that contains all of the information needed and
that includes all of the servers that host may need to contact for testing purposes. 3. Edit the /etc/nsswitch and change the "hosts" line to have "files" only.
Thanks, but that wasn't the question, and it raises its own difficult questions. I'll wait for Tim.
Tony Nelson wrote:
- Remove the /etc/resolv.conf
- Create a /etc/hosts file that contains all of the information needed and
that includes all of the servers that host may need to contact for testing purposes. 3. Edit the /etc/nsswitch and change the "hosts" line to have "files" only.
Thanks, but that wasn't the question, and it raises its own difficult questions. I'll wait for Tim.
Gee....
OK, the question was how can I set up a nameserver without knowing or wanting to know anything about a nameserver and without reading and understanding the documentation.
Then, yes, please wait for someone to do it for you.
It's nice to be held in such high regard, but I'm not the only one who plays with BIND...
On Thu, 2007-03-29 at 17:46 -0400, Tony Nelson wrote:
Tim -- is this also a solution for my problem? I have set up a local server (on CentOS 4.4) to test a replacement for a real server with a domain name (running RH 7.2). The local server should always resolve that domain name to itself, so as to properly test itself and not the real server.
Once you set up zone records on a machine, it'll use them instead of trying externally, as it already has an answer for queries (even if its a null answer). I do this for advert busting. I have a series of configuration entries for annoying domain names that'll return null answers for the network. That gets rid of various web browsing annoyances, centrally.
I added a series of lines like the following to my lan.conf file:
zone "adimages.com" { type master; file "dead.zone"; }; zone "admonitor.com" { type master; file "dead.zone"; }; zone "adsfac.net" { type master; file "dead.zone"; }; zone "advertising.com" { type master; file "dead.zone"; }; zone "amazingmedia.com" { type master; file "dead.zone"; };
Which causes any queries for those domains to get *MY* answer, not the one from their real master servers. The "dead.zone" file as as follows, it produces a "no answer" result, causing instant death for the attempt to browse to it.
$TTL 86400 @ IN SOA ns.localdomain. hostmaster.mail.localdomain. ( 200 ; serial 28800 ; refresh 7200 ; retry 604800 ; expire 86400 ; ttl )
IN NS ns.localdomain.
And that's the whole thing, there's no further entries in it. It works better than wildcarding, or playing with hosts files, as that directs queries to somewhere else, rather than aborting them.
The same applies if you provide real answers for a domain. They'll be used, instead of going out on the internet to get the records.
Just adding lines to /etc/hosts was insufficient, probably because DNS queries (dig, nslookup) still see the real server.
Various things will ignore the hosts file, I've it's best to avoid using it for all but the most basic of things (such as the machine's own addresses, if they're static, so it can run disconnected from the network without complaints).
I'm trying to avoid really understanding BIND at this time, as I have plenty of other issues to deal with, and the new server, when deployed, won't need this testing DNS configuration.
Understanding DNS is an important part of networking, understanding how BIND does its tricks, on top of that, is yet another thing worth finding out about.
Tim wrote:
Once you set up zone records on a machine, it'll use them instead of trying externally, as it already has an answer for queries (even if its a null answer). I do this for advert busting. I have a series of configuration entries for annoying domain names that'll return null answers for the network. That gets rid of various web browsing annoyances, centrally.
I added a series of lines like the following to my lan.conf file:
zone "adimages.com" { type master; file "dead.zone"; }; zone "admonitor.com" { type master; file "dead.zone"; }; zone "adsfac.net" { type master; file "dead.zone"; }; zone "advertising.com" { type master; file "dead.zone"; }; zone "amazingmedia.com" { type master; file "dead.zone"; };
Which causes any queries for those domains to get *MY* answer, not the one from their real master servers. The "dead.zone" file as as follows, it produces a "no answer" result, causing instant death for the attempt to browse to it.
$TTL 86400 @ IN SOA ns.localdomain. hostmaster.mail.localdomain. ( 200 ; serial 28800 ; refresh 7200 ; retry 604800 ; expire 86400 ; ttl )
IN NS ns.localdomain.And that's the whole thing, there's no further entries in it. It works better than wildcarding, or playing with hosts files, as that directs queries to somewhere else, rather than aborting them.
The same applies if you provide real answers for a domain. They'll be used, instead of going out on the internet to get the records.
I don't find a "lan.conf" file but I do have the following. Which one, if any, can I operate on to eliminate the loading of those darned links? My satellite connection has an inherent delay and waiting for six or eight ad links to be dealt with that I never view is maddening! The caching-nameserver works well enough once it looks up an address but there's still the time wasted collecting data that is never displayed, eliminated from Firefox by 'Adblock' and 'Flashblock.'
/etc/named.rfc1912.zones /usr/share/doc/bind-9.3.4/sample/etc/named.rfc1912.zones /usr/share/doc/bind-9.3.4/sample/var/named/localdomain.zone /usr/share/doc/bind-9.3.4/sample/var/named/localhost.zone /usr/share/doc/bind-9.3.4/sample/var/named/my.external.zone.db /usr/share/doc/bind-9.3.4/sample/var/named/my.internal.zone.db /usr/share/doc/bind-9.3.4/sample/var/named/slaves/my.ddns.internal.zone.db /usr/share/doc/bind-9.3.4/sample/var/named/slaves/my.slave.internal.zone.db /var/named/localdomain.zone /var/named/localhost.zone /var/named/chroot/etc/named.rfc1912.zones /var/named/chroot/var/named/localdomain.zone /var/named/chroot/var/named/localhost.zone
Thanks.
Bob Goodwin Zuni, Virginia
At 4:18 PM +0930 3/30/07, Tim wrote:
It's nice to be held in such high regard, but I'm not the only one who plays with BIND...
On Thu, 2007-03-29 at 17:46 -0400, Tony Nelson wrote:
Tim -- is this also a solution for my problem? I have set up a local server (on CentOS 4.4) to test a replacement for a real server with a domain name (running RH 7.2). The local server should always resolve that domain name to itself, so as to properly test itself and not the real server.
Once you set up zone records on a machine, it'll use them instead of trying externally, as it already has an answer for queries (even if its a null answer). I do this for advert busting. I have a series of configuration entries for annoying domain names that'll return null answers for the network. That gets rid of various web browsing annoyances, centrally.
I added a series of lines like the following to my lan.conf file:
zone "adimages.com" { type master; file "dead.zone"; }; zone "admonitor.com" { type master; file "dead.zone"; }; zone "adsfac.net" { type master; file "dead.zone"; }; zone "advertising.com" { type master; file "dead.zone"; }; zone "amazingmedia.com" { type master; file "dead.zone"; };
Which causes any queries for those domains to get *MY* answer, not the one from their real master servers. The "dead.zone" file as as follows, it produces a "no answer" result, causing instant death for the attempt to browse to it.
$TTL 86400 @ IN SOA ns.localdomain. hostmaster.mail.localdomain. ( 200 ; serial 28800 ; refresh 7200 ; retry 604800 ; expire 86400 ; ttl )
IN NS ns.localdomain.And that's the whole thing, there's no further entries in it. It works better than wildcarding, or playing with hosts files, as that directs queries to somewhere else, rather than aborting them.
The same applies if you provide real answers for a domain. They'll be used, instead of going out on the internet to get the records.
...
Thanks. I've installed bind and caching-nameserver. I put lines like
zone "mydomain.com" { type master; file "localhost.zone"; };
into named.conf and informed named with a kill SIGHUP. The computer gets its IP address via DHCP, so I added a line to /etc/dhclient-the0.conf
supersede domain-name-servers 127.0.0.1;
and then flapped the interface with
# ifdown eth0 ; ifup eth0
in order to get /etc/resolv.conf to use the caching nameserver. I was reminded that it is important to do both commands on one line if admistering "remotely". There might also be a simpler way to do it.
Named is working properly, but I'm a bit uncomfortable about not having defined any nameservers to use. Presumably it's just asking the root nameservers. I could hard-code what DHCP would return, but I wonder if there is a way to get that info into named automatically?
On Sat, 2007-03-31 at 13:31 -0400, Bob Goodwin - W2BOD wrote:
I don't find a "lan.conf" file but I do have the following. Which one, if any, can I operate on to eliminate the loading of those darned links?
Go back into this thread to my posting before that one, and see the two custom files of mine. That lan.conf is something I put in there, myself. The reason being that modifying original files can cause you problems when you update bind. Your modifications may disappear. Writing them externally means that the most you have to do to get them back is insert an include line, rather than redo all the other things you did.
With the current version of BIND, it looks like it doesn't come with the default named.conf file, and will start off using the other ones it comes with, if it doesn't find one. I put in a named.conf file, and make it refer to the other files (with an include).
On Sat, 2007-03-31 at 23:39 -0400, Tony Nelson wrote:
Named is working properly, but I'm a bit uncomfortable about not having defined any nameservers to use. Presumably it's just asking the root nameservers.
Yes, it's acting as a stand-alone name server. Probably the best way to use it, for simplicity's sake.
I could hard-code what DHCP would return, but I wonder if there is a way to get that info into named automatically?
I wouldn't bother, nor would I recommend changing named.conf all the time. If you really wanted, you could play with adding forwarders to your configuration file, either to always be used for looking up external addresses, or for when it can't find an answer by itself.
At 3:14 PM +0930 4/1/07, Tim wrote:
On Sat, 2007-03-31 at 23:39 -0400, Tony Nelson wrote:
Named is working properly, but I'm a bit uncomfortable about not having defined any nameservers to use. Presumably it's just asking the root nameservers.
Yes, it's acting as a stand-alone name server. Probably the best way to use it, for simplicity's sake.
Certainly it is good enough for the present use.
I could hard-code what DHCP would return, but I wonder if there is a way to get that info into named automatically?
I wouldn't bother, nor would I recommend changing named.conf all the time. If you really wanted, you could play with adding forwarders to your configuration file, either to always be used for looking up external addresses, or for when it can't find an answer by itself.
OK. Maybe when I understand about forwarders.