Node: dns.js not properly iterating through DNS array

[MatthewPinheiro]

• Version: 8.11.3
• Platform: Windows 10 Pro v1709
• Subsystem: dns

I'll preface this by saying that there's a significant chance I'm simply misunderstanding dns.setServers, but that being said, I believe I'm experiencing undesired functionality.

Given var dnsServer1 and var dnsServer2 are two valid IP Address strings of two valid DNS servers, and given that var host is a valid string which resolves to an IP on dnsServer1 but not on dnsServer2, consider the following code:

dns.setServers([dnsServer2, dnsServer1]);
dns.resolve(host, (err, records) => {
    if (err) {
        console.log(err);
    } else {
        console.log(records);
});

If the above is run, an ENOTFOUND error will be produced and records will be undefined. However, if the order of the array passed into dns.setServers is reversed, i.e.

dns.setServers([dnsServer1, dnsServer2])

and the same call to dns.resolve is made, no error will be produced and records will have resolved to a valid IP.

Isn't dns.resolve supposed to check through each DNS IP set by dns.setServers until either one is able to resolve host or none of them are able to do so? Here it instead seems the entire resolve operation fails as soon as the first DNS IP fails, and succeeds as soon as the first DNS IP succeeds.

I've tried to find more thorough documentation but couldn't find anything that said one way or the other.

[bnoordhuis]

Isn't dns.resolve supposed to check through each DNS IP set by dns.setServers until either one is able to resolve host or none of them are able to do so?

No, the second one is the fallback for when the first one times out or errors; not when it reports "no such host exists."

I've tried to find more thorough documentation but couldn't find anything that said one way or the other.

It's basically how /etc/resolv.conf works. I suppose a note could be added to the docs. PR welcome.

[BeniCheni]

Hello, just minor friendly message that I'm happy to give this "_good first issue_" issue an attempt. Hopefully I could make progress to open a PR soon.

P.S.: Updated below to strike through my outdated comments from previous edits:
Noticed that this #21273 back-filled a contributor to the README. If I open a PR for this issue, I would also update the "collaborators" field to include self, since I've landed 2 "_good first issue_" PRs before. (#20092 & #20460).

I'd hope it's okay to add myself as a "contributor". But if my assumption is incorrect about the definition of the "contributor", please kindly let me know. (So I won't update the "contributors" field of README mistakenly.)

P.S.: Question - What does Collaborator Emeriti imply for the context of contributors?

(Sorry that my question is not related to this issue; just got curious about the definition of the keyword "_Emeriti_"; I can also do a search of the keyword in the docs in the repo to try to learn more about that)

[richardlau]

@BeniCheni Collaborators are not the same as contributors. Please see https://github.com/nodejs/node/blob/master/GOVERNANCE.md for more information about collaborators.

All contributors to Node.js are listed in https://github.com/nodejs/node/blob/master/AUTHORS (I believe https://github.com/nodejs/node/blob/4970e2bec8a15739739ac4317158233c8cf32c6d/AUTHORS#L2190 is you).

[Shivang44]

I didn't see any PR for 4 days from the user above so I made one: #21469

Let me know if everyone likes the note. Perhaps we could clarify further on what errors it will continue checking other servers? I couldn't find this information in the documentation so I'm not sure.

[MatthewPinheiro]

Thanks for making the PR @Shivang44. Didn't mean to completely abandon my own issue, but just completely forgot to make a PR once I was home. Thanks!

[Shivang44]

Landed in 95205a6. I believe this issue can be closed as long as @MatthewPinheiro is okay with it