Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Some of my domains run using two different IPs for each ns property in their DNS records, e.g.:

ns1.example-1.com 5.5.5.10
ns2.example-1.com 5.5.5.11

and some of my domains have the same IP for all, e.g.:

ns1.example-2.com 5.5.5.10
ns2.example-2.com 5.5.5.10

Despite of the fact that the examples above work identically well, if I check-mark Test DNS Servers at registar's domain manager using same IP for each ns1 and ns2 (see example-2.com), then the error is returned, requiring to have two different IPs for each ns property? (see example-1.com)

My question is straightforward:
Why do you need two different IPs for each ns record in order to run DNS Servers' Test but at the same time if you just set records without testing procedure, then the registar adds them as it is with no problem?

share|improve this question
up vote 1 down vote accepted

It's most likely because of RFC 1304 (DOMAIN NAMES - CONCEPTS AND FACILITIES), and your domain registrar is probably attempting to force you to abide by it in their test system (but apparently not in their production side).

From RFC 1304, section 4.1 (Name Servers - Introduction):

A given zone will be available from several name servers to insure its availability in spite of host or communication link failure. By administrative fiat, we require every zone to be available on at least two servers, and many zones have more redundancy than that.

Also check out:

share|improve this answer
    
thank you for your answer with interesting links! Can you tell me, what you think is the point of having 2 IPs from the same network? For example, 5.3.3.10 and 5.3.3.11? If first one fails to work, then the second one will most likely also fail as it's the same hardware serving them? More reasonable would be to have IPs from different sub-networks? – Ilia Rostovtsev Mar 23 '13 at 4:53

NS1 means name server 1. This is the first name server that is attempted for resolution. NS2 would mean name server 2. It is the second name server that is attempted when ns1 is unresponsive. By setting ns1 and ns2 to the same IP address you are removing redundancy. It will work fine, unless the name server at the IP address for NS1 becomes unresponsive. Then only the entries with a different IP address for NS2 will continue to work (unless it too is down). I've seen people have up to 6 backup name servers (I currently use 3).

The test procedure likely includes making sure that the IP addresses are not the same, and returns an error if they are, because it destroys redundancy.

share|improve this answer
    
thanks for your answer! Unfortunately the second name server is NOT attempted when the first one is unresponsive! The are queried in arbitrary way! I personally can't understand the reason for having two different IPs unless they are from the different sub-networks? – Ilia Rostovtsev Mar 23 '13 at 4:48
    
This is incorrect. The names given in the rdata portions of NS records are just names. They do not indicate an order in which to query nameservers (in fact, they do not have to be of the form ns<number>) or carry any other semantic meaning. – Blacklight Shining Feb 5 at 13:52
    
I checked the RFC. NS falls under a QDATA type. So the fact that the title of the IP address is NS appears to be relevant (that's what the RFC says). Now, does the number (the one in NS1) matter? You have two Names Servers listed NS1 and NS2. How is one selected over the other? Are you asserting it's random? Is there another mechanism used? If you're going to point out I'm wrong, that's great, but it's only worth something if you include what makes it right. – Everett Feb 11 at 17:59

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.