Re: [SLUG] DHCP - DDNS not updating
(Hi I'll reply to this thread as there seems to be a couple of threads going on the same subject) I've had a very similar setup to you in the past. I never had this much trouble. I had only centos 5.x servers. You've got split DNS. Internally, DNS resolves to your internal DNS server from your clients and you can see what you have allowed from internal as it's all internal. Externally, the world goes to dnsmadeeasy.com as your domains are delegated to this (i'm presuming). So if for example you are hosting a webserver, then dnsmadeeasy would point that domain to your MODEM external ip address. On your modem you would have a virtual server setup with port 80 forwarded to your internal webserver ip address. So from internal you get to the clients webserver from your internal DNS. From external you get to your clients webserver from being redirected through dnsmadeeasy to your external IP of your modem and then from your modem forwarded to your webserver. You could even have your internal clients all point to the modem for DNS. The modem would ask dnsmadeeasy where that domain was, it would point to your modem and then the modem would port forward them all back into your internal webserver. I never had any issues with dns key files or dnssec or whatever as I never needed to use it/them. I would recommend removing all these until the basics are working solidly. Have you tried using webmin to setup dns on your internal dns server? Thanks, Ben Donohue On 17/02/2011 3:16 PM, Kyle wrote: Peter, exactly!! THAT IS MY ISSUE I believe. But I have not yet found a log to give me sufficient info to nut out WHY. All my config files are presently up for the world to see at; https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30159 And from what I've read (LOTS in the last couple of days), they're picture perfect. Kind Regards Kyle On 17/02/11 3:02 PM, pe...@chubb.wattle.id.au wrote: I strongly suspect that the key setup is incorrect. .it will fail because of an authorisation problem. Peter C -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
On Tue, Feb 15, 2011 at 07:56:55PM +1100, Kyle wrote: Sorry for the late reply; I've been busy. I have always used BIND with rndc.key and it used to work. What's then the difference between nsupdate and rndc and using BIND? They have two quite different functions. nsupdate is used to modify zone data by sending dynamic DNS updates. rndc is used to control the name server itself, for example, to stop or restart the server, to reload config and/or zone files. John -- Yes, but then you'd have a computer running OS X. There may be an upside, but I'm not sure what it is. -- Matt Palmer -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Ok, thanks John. I've tried lots since this thread started to the extent I installed a whole fresh machine on 192 subnet only, skimmed dhcpd and named confs down to a simple, by the book, 1 domain setup and I still get the same problem even on the fresh host. And this is a CentOS 5.5 (final) install. I.e. the latest they publish. Reply from update query: ;; -HEADER- opcode: UPDATE, status: _*NOTAUTH*_, id: 39734 ;; flags: qr ra ; ZONE: 0, PREREQ: 0, UPDATE: 0, ADDITIONAL: 1 ;; TSIG PSEUDOSECTION: domain1.com.0ANYTSIGhmac-md5.sig-alg.reg.int. 1297906252 300 16 someSecretHashHere 39734 NOERROR 0 If I'm reserving an IP for a specific host in dhcpd.conf, am I supposed to then be already placing a PTR record in the reverse zone file for the reservation? If so, doesn't that simply defeat the whole purpose of dhcp? Kind Regards Kyle On 17/02/11 1:26 PM, John Clarke wrote: On Tue, Feb 15, 2011 at 07:56:55PM +1100, Kyle wrote: Sorry for the late reply; I've been busy. I have always used BIND with rndc.key and it used to work. What's then the difference between nsupdate and rndc and using BIND? They have two quite different functions. nsupdate is used to modify zone data by sending dynamic DNS updates. rndc is used to control the name server itself, for example, to stop or restart the server, to reload config and/or zone files. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
On Thu, Feb 17, 2011 at 01:38:58PM +1100, Kyle wrote: I've tried lots since this thread started to the extent I installed a whole fresh machine on 192 subnet only, skimmed dhcpd and named confs down to a simple, by the book, 1 domain setup and I still get the same problem even on the fresh host. And this is a CentOS 5.5 (final) install. I.e. the latest they publish. Are you having problems updating the A record or the PTR record, or both? Can you manually update either or both with nsupdate run on the DNS server. Can you do the same with nsupdate on the DHCP server? Are the clocks on the DHCP server and the DNS server synchronised? I don't remember why this mattered (it's been years since I set this up myself), but I do remember having updates fail when the clocks were not synchronised. Can you send me your DNS and DHCP config files? If I'm reserving an IP for a specific host in dhcpd.conf, am I supposed to then be already placing a PTR record in the reverse zone file for the reservation? By reserving, do you mean that you've configured your DHCP server to allocate fixed addresses based on the host's MAC address? I wouldn't have thought it would matter, that the DHCP server would update both the A and PTR records anyway, as long as you've configured it to update both zones in dhcpd.conf. If so, doesn't that simply defeat the whole purpose of dhcp? No. The purpose of DHCP is to allow hosts to get an IP address and other information without having to manually configure each host. The DDNS update is just a bonus :-) John -- What's the use of having a cupboard full of dingy little machines, none of which can run continuously for more than five minutes without needing 15 minutes of cooldown? Our NT admins have an answer to that. -- Joe Moore -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
John, Answers inline. Kind Regards Kyle On 17/02/11 2:00 PM, John Clarke wrote: Are you having problems updating the A record or the PTR record, or both? ## Both. DHCP is simply not updating DNS journals when a lease is renewed. Can you manually update either or both with nsupdate run on the DNS server. ## No, nsupdate extract from last email was run on the dns server. Can you do the same with nsupdate on the DHCP server? ## Same box. Are the clocks on the DHCP server and the DNS server synchronised? I don't remember why this mattered (it's been years since I set this up myself), but I do remember having updates fail when the clocks were not synchronised. ## Same box Can you send me your DNS and DHCP config files? ## Whole sordid sorry saga here (files and all), fresh this a.m.; https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30159 -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Kyle == Kyle k...@attitia.com writes: Kyle John, Answers inline. Kyle Kyle Kind Regards Kyle Kyle Kyle On 17/02/11 2:00 PM, John Clarke wrote: Are you having problems updating the A record or the PTR record, or both? Kyle ## Both. DHCP is simply not updating DNS journals when a lease Kyle is renewed. I strongly suspect that the key setup is incorrect. Take a look in /etc/bind/your-key-file It should look like key DHCP_updater { algorithm hmac-md5; secret longstringofcharactersendingin==; }; (this is the one referred to in /etc/dhcp3/dhcpd.conf with include /etc/bind/your_key_file ) And in /etc/bind/named.conf you should have: include /etc/bind/your_key_file and controls { inet 127.0.0.1 port 953 allow { 127.0.0.1;} keys { DHCP_updater; ... }; }; And again in each zone statement: zone domain.com in { type master; file /etc/bind/domain.zone; allow-update { key DHCP_updater; }; }; or similar. If the name of the key isn't the same, or you're including different keyfiles, it will fail because of an authorisation problem. Peter C -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Peter, exactly!! THAT IS MY ISSUE I believe. But I have not yet found a log to give me sufficient info to nut out WHY. All my config files are presently up for the world to see at; https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30159 And from what I've read (LOTS in the last couple of days), they're picture perfect. Kind Regards Kyle On 17/02/11 3:02 PM, pe...@chubb.wattle.id.au wrote: I strongly suspect that the key setup is incorrect. .it will fail because of an authorisation problem. Peter C -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
On Thu, Feb 17, 2011 at 02:48:45PM +1100, Kyle wrote: Can you manually update either or both with nsupdate run on the DNS server. ## No, nsupdate extract from last email was run on the dns server. OK, so if this doesn't work, then the DHCP server won't be able to do it either. You need to fix this problem before you can get anything else working. What name did you give the key when you created it with dnssec-keygen? Was it domain1.com? Is that what name you passed to nsupdate (in the argument to -y)? Do you still have the keyfiles generated by dnssec-keygen? Does nsupdate work if you use -k instead of -y to pass the key? Does rndc status work? rndc reload? John -- ... I can use the cheaper frozen goat for IDE? Though you _can_ sacrifice things to an IDE bus, there are no reports of them ever sucking any less as a result, so it's likely a waste of a goat. -- Anthony de Boer -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
John, 'domain1.com' is obfuscated from the real value. But rest assured I am being painstakingly anal in ensuring the values are the same including the 'key name' in named and dhcpd being exactly the same as used in the dnssec-keygen command. transcript of interaction. Long story short, the error is 'NOTAUTH'. And the result is the same no matter whether I use the .private or .key files. [root@server3 etc]# rndc status number of zones: 8 debug level: 0 xfers running: 0 xfers deferred: 0 soa queries in progress: 0 query logging is OFF recursive clients: 0/1000 tcp clients: 0/100 server is up and running [root@server3 etc]# rndc reload server reload successful [root@server3 etc]# nsupdate -d -k Kdomain1.com.+157+63230.private Creating key... server 127.0.0.1 zone domain1.com update add client1.domain1.com 86400 A 192.168.1.100 send Sending update to 127.0.0.1#53 Outgoing update query: ;; -HEADER- opcode: UPDATE, status: NOERROR, id: 2442 ;; flags: ; ZONE: 1, PREREQ: 0, UPDATE: 1, ADDITIONAL: 1 ;; ZONE SECTION: ;domain1.com.INSOA ;; UPDATE SECTION: client1.domain1.com.86400INA192.168.1.100 ;; TSIG PSEUDOSECTION: domain1.com.0ANYTSIGhmac-md5.sig-alg.reg.int. 1297920682 300 16 someSecretHere 2442 NOERROR 0 Reply from update query: ;; -HEADER- opcode: UPDATE, status: NOTAUTH, id: 2442 ;; flags: qr ra ; ZONE: 0, PREREQ: 0, UPDATE: 0, ADDITIONAL: 1 ;; TSIG PSEUDOSECTION: domain1.com.0ANYTSIGhmac-md5.sig-alg.reg.int. 1297920682 300 16 anotherSecretHere 2442 NOERROR 0 quit Kind Regards Kyle On 17/02/11 3:42 PM, John Clarke wrote: On Thu, Feb 17, 2011 at 02:48:45PM +1100, Kyle wrote: What name did you give the key when you created it with dnssec-keygen? Was it domain1.com? Is that what name you passed to nsupdate (in the argument to -y)? Do you still have the keyfiles generated by dnssec-keygen? Does nsupdate work if you use -k instead of -y to pass the key? Does rndc status work? rndc reload? -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
On Thu, Feb 17, 2011 at 04:37:48PM +1100, Kyle wrote: 'domain1.com' is obfuscated from the real value. But rest assured I am being painstakingly anal in ensuring the values are the same including the 'key name' in named and dhcpd being exactly the same as used in the dnssec-keygen command. OK, I just wanted to be sure, because the only way I've been able to reproduce similar symptoms to yours was by using a different name. [root@server3 etc]# rndc reload server reload successful I thought this might provide a clue, but I've tested it on my server here and rndc seems to work even if the key it's told to use is not authorised. Oh well. Reply from update query: ;; -HEADER- opcode: UPDATE, status: NOTAUTH, id: 2442 ;; flags: qr ra ; ZONE: 0, PREREQ: 0, UPDATE: 0, ADDITIONAL: 1 ;; TSIG PSEUDOSECTION: domain1.com.0ANYTSIGhmac-md5.sig-alg.reg.int. 1297920682 300 16 anotherSecretHere 2442 NOERROR 0 Now this is slightly different to anything I've been able to reproduce. If I give it the wrong key I see BADKEY in that last line instead of NOERROR. This is just a guess because I've pretty much hit the limits of my knowledge, and I've never used BIND's views, but could it be something to do with the different views you've configured? You're trying to do the update from localhost, so that matches the view localhost_resolver, but updates aren't allowed in that view configuration. Updates are allowed in the view internal, which also matches localhost, but I wonder if BIND is simply using the first match and thus disallowing updates? John -- Vs lbh'er ernqvat guvf, lbh ernyyl bhtug gb trg bhg zber -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Ah, ok. Thanks for the explanation John. I have never before had to use nsupdate. I just tried it because Peter suggested it and I figured it's a way to test dns updates manually. I have always used BIND with rndc.key and it used to work. What's then the difference between nsupdate and rndc and using BIND? Kind Regards Kyle On 15/02/11 6:52 PM, John Clarke wrote: On Tue, Feb 15, 2011 at 05:35:10PM +1100, Kyle wrote: domain domain1.com incorrect section name: domain I suspect you mean zone domain1.com. domain is not a valid command. nsupdate -k /etc/rndc.key - The man page says that that format requires a filename in the format 'K{name}.+157.+{random}.private'. That's a new one on me. Where, why how is that needed? That's been the case for as long as I've been using nsupdate, at least five years. The filename format is what dnssec-keygen outputs when you ask it to generate a key. One other thing you need to make sure of is that the client and server have their clocks synchronised (e.g. with ntp), otherwise the update will fail. There's an nsupdate HOWTO here: http://caunter.ca/nsupdate.txt and I have a page explaining how to get DHCP3 to do DDNS updates here: http://kirriwa.net/john/doc/ddns.html John -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Kyle == Kyle k...@attitia.com writes: Kyle Ah, ok. Thanks for the explanation John. Kyle I have never before had to use nsupdate. I just tried it because Kyle Peter suggested it and I figured it's a way to test dns updates Kyle manually. Yes --- it'll check that the key you have actually works, and that the nameserver allows DDNS updates. The problem you describe has three components it'd be good to try testing them one at a time. 1. Is named accepting DDNS updates from the machine running DHCPD, using the same key as DHCPD? 2. Is DHCPD generating DDNS requests for the correct zone? 3. Are clients getting the right address/name combinations? Peter C Kyle I have always used BIND with rndc.key and it used to Kyle work. What's then the difference between nsupdate and rndc and Kyle using BIND? Kyle Kyle Kind Regards Kyle Kyle Kyle On 15/02/11 6:52 PM, John Clarke wrote: On Tue, Feb 15, 2011 at 05:35:10PM +1100, Kyle wrote: domain domain1.com incorrect section name: domain I suspect you mean zone domain1.com. domain is not a valid command. nsupdate -k /etc/rndc.key - The man page says that that format requires a filename in the format 'K{name}.+157.+{random}.private'. That's a new one on me. Where, why how is that needed? That's been the case for as long as I've been using nsupdate, at least five years. The filename format is what dnssec-keygen outputs when you ask it to generate a key. One other thing you need to make sure of is that the client and server have their clocks synchronised (e.g. with ntp), otherwise the update will fail. There's an nsupdate HOWTO here: http://caunter.ca/nsupdate.txt and I have a page explaining how to get DHCP3 to do DDNS updates here: http://kirriwa.net/john/doc/ddns.html John Kyle -- SLUG - Sydney Linux User's Group Mailing List - Kyle http://slug.org.au/ Subscription info and FAQs: Kyle http://slug.org.au/faq/mailinglists.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
I reckon I've narrowed it down to the fact that the DNS server is not accepting the key it's being offered. The final transaction I receive from an nsupdate exchange offers; --- ;; TSIG PSEUDOSECTION: domain1_key.0ANYTSIGhmac-md5.sig-alg.reg.int. 1297798631 300 16 P0HVLAVNuSartbLEAqQ3Lg== 22243 NOERROR 0 Reply from update query: ;; -HEADER- opcode: UPDATE, status: _*NOTAUTH*_, id: 22243 ;; flags: qr ra ; ZONE: 0, PREREQ: 0, UPDATE: 0, ADDITIONAL: 1 ;; TSIG PSEUDOSECTION: domain1_key.0ANYTSIGhmac-md5.sig-alg.reg.int. 1297798631 300 16 8KsZzvC3KJQNnX4EErqdXQ== 22243 NOERROR 0 -- So pretty much the same error as syslog was offering up. But I don't seem to be able to hit on the right google keyword combination to offer up the solution. Can anyone with some in-depth (basically more than me - which is not going to be hard) knowledge offer me avenues to chase down? I have checked, dbl-checked and triple-checked that the key is generated correctly. I have the correct key notation in named.conf, dhcpd.conf, rndc.conf and rndc.key. But I'm obviously missing come pivotal step in either permissions or configuration. Kind Regards Kyle -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] DHCP - DDNS not updating
Hi SLUG, I have an issue that is becoming more annoying the more I try to track it down. Can I ask those of you interested to take a look at the following thread please: https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30080forum=40 I realise this might be a little brave ... to ask you to go outside the list, but it is a fairly detailed thread (i.e. lengthy read) with all the necessary info on the box's config and all the testing I have done. In short; --- mixed clients XP, 7, OSX, Ubuntu 10 to a CentOS 5.5 svr all in an MS WORKGROUP type setup. DHCP licences handed out, some reserved. DNS zones supposedly set up to provide name-based network browsing within the WORKGROUP domain and for internal clients to find inbound mail servers for 5 other domains. --- But try as I might, I can't seem to get the DNS updated anymore and I swear it USED TO WORK. The only thing I have since done is update the packages as they came out. I am as far as the error: Unable to add forward map from client1.domain1.com to 192.168.1.104: not authorized and that's where it stops. No amount of googlage has so far shone a torch on the issue. The disheartening thing is the bloody redmond boxes CAN find other hosts by name, but not the linux or OSX boxes. If you have the time and incl. to have a read and point out where the hell I'm going wrong pls, I'd be grateful. -- Kind Regards Kyle -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Hi, I'd check that you are port forwarding or virtual server from the modem port 53 tcp and udp to the internal dns server reboot everything starting from the modem and working your way inwards try to simplify the problem. Run up a new box if you can and stick it temporarily on the inside of the modem and see if it is getting dns correctly... go to websites etc if working then place it on the other side of the svr1 and test from there... Thanks, Ben Donohue On 15/02/2011 11:59 AM, Kyle wrote: Hi SLUG, I have an issue that is becoming more annoying the more I try to track it down. Can I ask those of you interested to take a look at the following thread please: https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30080forum=40 I realise this might be a little brave ... to ask you to go outside the list, but it is a fairly detailed thread (i.e. lengthy read) with all the necessary info on the box's config and all the testing I have done. In short; --- mixed clients XP, 7, OSX, Ubuntu 10 to a CentOS 5.5 svr all in an MS WORKGROUP type setup. DHCP licences handed out, some reserved. DNS zones supposedly set up to provide name-based network browsing within the WORKGROUP domain and for internal clients to find inbound mail servers for 5 other domains. --- But try as I might, I can't seem to get the DNS updated anymore and I swear it USED TO WORK. The only thing I have since done is update the packages as they came out. I am as far as the error: Unable to add forward map from client1.domain1.com to 192.168.1.104: not authorized and that's where it stops. No amount of googlage has so far shone a torch on the issue. The disheartening thing is the bloody redmond boxes CAN find other hosts by name, but not the linux or OSX boxes. If you have the time and incl. to have a read and point out where the hell I'm going wrong pls, I'd be grateful. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Ben, thanks for taking the time to look. Perhaps I wasn't clear. My issues are not external clients not updating; rather internal only. I.e. Everything behind my firewalls. Kind Regards Kyle On 15/02/11 1:11 PM, Ben Donohue wrote: Hi, I'd check that you are port forwarding or virtual server from the modem port 53 tcp and udp to the internal dns server reboot everything starting from the modem and working your way inwards try to simplify the problem. Run up a new box if you can and stick it temporarily on the inside of the modem and see if it is getting dns correctly... go to websites etc if working then place it on the other side of the svr1 and test from there... Thanks, Ben Donohue On 15/02/2011 11:59 AM, Kyle wrote: Hi SLUG, I have an issue that is becoming more annoying the more I try to track it down. Can I ask those of you interested to take a look at the following thread please: https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30080forum=40 I realise this might be a little brave ... to ask you to go outside the list, but it is a fairly detailed thread (i.e. lengthy read) with all the necessary info on the box's config and all the testing I have done. In short; --- mixed clients XP, 7, OSX, Ubuntu 10 to a CentOS 5.5 svr all in an MS WORKGROUP type setup. DHCP licences handed out, some reserved. DNS zones supposedly set up to provide name-based network browsing within the WORKGROUP domain and for internal clients to find inbound mail servers for 5 other domains. --- But try as I might, I can't seem to get the DNS updated anymore and I swear it USED TO WORK. The only thing I have since done is update the packages as they came out. I am as far as the error: Unable to add forward map from client1.domain1.com to 192.168.1.104: not authorized and that's where it stops. No amount of googlage has so far shone a torch on the issue. The disheartening thing is the bloody redmond boxes CAN find other hosts by name, but not the linux or OSX boxes. If you have the time and incl. to have a read and point out where the hell I'm going wrong pls, I'd be grateful. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Hi Kyle, I suggest looking in /var/log/messages (I believe the default confiuration for CENTOS logs everything into there ) You should see lines like: Feb 15 06:25:34 vampire dhcpd: DHCPREQUEST for A.B.C.D from some:mac:address:here via eth0 then you might see something like Feb 15 06:25:35 vampire named[3921]: client 127.0.0.1#53: update 'keg.ertos.in.nicta.com.au/IN' denied or, Feb 14 17:35:04 vampire named[4854]: client 127.0.0.1#42217: updating zone 'keg.ertos.in.nicta.com.au/IN': update unsuccessful: bigmac.keg.ertos.in.nicta.com.au: 'name not in use' prerequisite not satisfied (YXDOMAIN) You may also want to check that can update named manually, using nsupdate. As root, on the machine running dhcpd, do nsupdate -k keyfile server whateveryournameserveris domain domain.com update add fred.domain.com 864000 A 192.168.1.100 show send or similar. Then try host fred.domain.com to see if you can see the record you just added. Peter C -- Dr Peter Chubb peter DOT chubb AT nicta.com.au http://www.ertos.nicta.com.au ERTOS within National ICT Australia All things shall perish from under the sky/Music alone shall live, never to die -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Ben, Ok, maybe I'm misunderstanding you. Why would I need to be port forwarding? All my relevant hosts and my DNS server all sit on the 192.168 subnet all behind the same firewall with no reason to go near the modem? The DNS server does act as a firewall, but yes, the relevant ports on the eth1 side for DNS are open (namely 53 for whatever reason - can't remember now - 953). For clarification the setup is modem (FW) | Server (FW, DNS, DHCP, etc) eth0 = 10.X subneteth1 = 192.168 subnet DHCP listens on 192.168.x.x | internal DHCP hosts on 192.168 subnet Kind Regards Kyle On 15/02/11 1:47 PM, Ben Donohue wrote: yes that's what I'm talking about. are you port forwarding DNS to your internal dns server? setup a test box on the inside of the modem and see if it works with just the modem in place. Thanks, Ben Donohue On 15/02/2011 1:28 PM, Kyle wrote: Ben, thanks for taking the time to look. Perhaps I wasn't clear. My issues are not external clients not updating; rather internal only. I.e. Everything behind my firewalls. Kind Regards Kyle On 15/02/11 1:11 PM, Ben Donohue wrote: Hi, I'd check that you are port forwarding or virtual server from the modem port 53 tcp and udp to the internal dns server reboot everything starting from the modem and working your way inwards try to simplify the problem. Run up a new box if you can and stick it temporarily on the inside of the modem and see if it is getting dns correctly... go to websites etc if working then place it on the other side of the svr1 and test from there... Thanks, Ben Donohue On 15/02/2011 11:59 AM, Kyle wrote: Hi SLUG, I have an issue that is becoming more annoying the more I try to track it down. Can I ask those of you interested to take a look at the following thread please: https://www.centos.org/modules/newbb/viewtopic.php?topic_id=30080forum=40 I realise this might be a little brave ... to ask you to go outside the list, but it is a fairly detailed thread (i.e. lengthy read) with all the necessary info on the box's config and all the testing I have done. In short; --- mixed clients XP, 7, OSX, Ubuntu 10 to a CentOS 5.5 svr all in an MS WORKGROUP type setup. DHCP licences handed out, some reserved. DNS zones supposedly set up to provide name-based network browsing within the WORKGROUP domain and for internal clients to find inbound mail servers for 5 other domains. --- But try as I might, I can't seem to get the DNS updated anymore and I swear it USED TO WORK. The only thing I have since done is update the packages as they came out. I am as far as the error: Unable to add forward map from client1.domain1.com to 192.168.1.104: not authorized and that's where it stops. No amount of googlage has so far shone a torch on the issue. The disheartening thing is the bloody redmond boxes CAN find other hosts by name, but not the linux or OSX boxes. If you have the time and incl. to have a read and point out where the hell I'm going wrong pls, I'd be grateful. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
Ok, that get's more info. - [root@server1 ~]# nsupdate -k /etc/rndc.key could not read key from /etc/rndc.key: unexpected token [root@server1 ~]# man nsupdate [root@server1 ~]# nsupdate -y domain1_key:theSecretPhraseFrom_rndc.key server 127.0.0.1 domain domain1.com incorrect section name: domain update add client1.domain1.com 864000 A 192.168.1.104 show Outgoing update query: ;; -HEADER- opcode: UPDATE, status: NOERROR, id: 0 ;; flags: ; ZONE: 0, PREREQ: 0, UPDATE: 0, ADDITIONAL: 0 ;; UPDATE SECTION: client1.domain1.com.864000INA192.168.1.104 send could not find enclosing zone [root@server1 ~]# host client1 Host client1 not found: 3(NXDOMAIN) [root@server1 ~]# host client1.domain1.com Host client1.domain1.com not found: 3(NXDOMAIN) --- So questions arising from that are; nsupdate -k /etc/rndc.key - The man page says that that format requires a filename in the format 'K{name}.+157.+{random}.private'. That's a new one on me. Where, why how is that needed? Just that error alone leads me to think the not authorized part of the 'forward map' error I was getting has to do with not being able to find the key. But then, after issuing the send command, I get the error; could not find enclosing zone. I can imagine if the key were not being found, the zone would not be found but don't know how/where that might becoming from? I'm going to assume the; domain domain1.com incorrect section name: domain is just a spurious section because the domain domainName command doesn't exist. Kind Regards Kyle On 15/02/11 1:55 PM, Peter Chubb wrote: Hi Kyle, I suggest looking in /var/log/messages (I believe the default confiuration for CENTOS logs everything into there ) You should see lines like: Feb 15 06:25:34 vampire dhcpd: DHCPREQUEST for A.B.C.D from some:mac:address:here via eth0 then you might see something like Feb 15 06:25:35 vampire named[3921]: client 127.0.0.1#53: update 'keg.ertos.in.nicta.com.au/IN' denied or, Feb 14 17:35:04 vampire named[4854]: client 127.0.0.1#42217: updating zone 'keg.ertos.in.nicta.com.au/IN': update unsuccessful: bigmac.keg.ertos.in.nicta.com.au: 'name not in use' prerequisite not satisfied (YXDOMAIN) You may also want to check that can update named manually, using nsupdate. As root, on the machine running dhcpd, do nsupdate -k keyfile server whateveryournameserveris domain domain.com update add fred.domain.com 864000 A 192.168.1.100 show send or similar. Then try host fred.domain.com to see if you can see the record you just added. Peter C -- Dr Peter Chubb peter DOT chubb AT nicta.com.au http://www.ertos.nicta.com.au ERTOS within National ICT Australia All things shall perish from under the sky/Music alone shall live, never to die -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
On Tue, Feb 15, 2011 at 05:35:10PM +1100, Kyle wrote: domain domain1.com incorrect section name: domain I suspect you mean zone domain1.com. domain is not a valid command. nsupdate -k /etc/rndc.key - The man page says that that format requires a filename in the format 'K{name}.+157.+{random}.private'. That's a new one on me. Where, why how is that needed? That's been the case for as long as I've been using nsupdate, at least five years. The filename format is what dnssec-keygen outputs when you ask it to generate a key. One other thing you need to make sure of is that the client and server have their clocks synchronised (e.g. with ntp), otherwise the update will fail. There's an nsupdate HOWTO here: http://caunter.ca/nsupdate.txt and I have a page explaining how to get DHCP3 to do DDNS updates here: http://kirriwa.net/john/doc/ddns.html John -- Bloody kids. If it hasn't got a front panel and you don't have to toggle in the bootstrap by hand you haven't experienced a real installation. -- Peter da Silva -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] DHCP - DDNS not updating
On Tue, Feb 15, 2011 at 05:13:05PM +1100, Kyle wrote: All my relevant hosts and my DNS server all sit on the 192.168 subnet all behind the same firewall with no reason to go near the modem? The DNS server does act as a firewall, but yes, the relevant ports on the eth1 side for DNS are open (namely 53 for whatever reason - can't remember now - 953). You need port 53 TCP and UDP, and port 953 TCP. Port 953 is bind's default control port, i.e. what rndc uses to talk to the server. nsupdate uses TCP port 53 IIRC. John -- Active ECM, coming soon to a WLAN near you! -- Chris Suslowicz -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html