Questions[resolve](server)

We have a lot of users on our VoIP software. The softphone checks each user's domain every minute, users make calls, send messages, etc. the usual VoIP stuff. Each user usually makes DNS checks several times a minute and altogether there are quite a few requests.

We cannot control to which public DNS the end-user tries resolving. When it is to Google public DNS sometimes, randomly it fails. To others like Cloudflare (1.1.1.1) it always works.

We've researched this and the issue is not the amount of DNS checks from the end-user towards Google DNS, which we know has a limit. Actually, the issue is from Google DNS to our Name Server when Google refuses to check the subdomain SOMETIMES. It seems that there is another limit here.

Here is an example of an unsuccessful request towards 8.8.8.8 after a nslookup (the domain exists).

dns.google can't find 7157599388.sip.teltel.io: Non-existent domain

Any idea how to change this limit?

Help will be highly appreciated!

Thank you in advance!

Domain name: abc.com Name Server: ns1.abc.com and ns2.abc.com

I am running a DNS server and hosting the zone (abc.com) on-premise.

When a client wants to resolve *.abc.com, it will query ns1.abc.com or ns2.abc.com, that is straight forward.

The question is how does it resolve ns1.abc.com before it can sends query to it?

Having nginx installed and Apache on port 8080. The trouble is: I'm building site for production domain (let's say mysite.com) and have everything configured for that domain. I change my Windows hosts file to redirect mysite.com to server's IP and it worked well... before I installed nginx. Nginx resolves domains itself and (obviously) can't resolve the domain mysite.com

Log says:

2012/02/07 07:35:29 [error] 11758#0: *1 mysite.com could not be resolved (110: Operation timed out), client: 89.112.11.xx, server: myssite.com, request: "GET / HTTP/1.1", host: "mysite.com"

Well, then I try to change hosts file on my server. I do this thing in the /etc/hosts:

127.0.0.1 mysite.com

Okay, let's try! If this took effect, at least wget should find it correct, right? As far as I have Apache on the 8080th port, i can wget the main page:

wget mysite.com

It downloads the correct page. This means that /etc/hosts does well.

But ngnix still does not resolve it! Even after I restarted the server. Why? Does it have its own resolver or what? How to make it work?

Assuming my server is www.mydomain.com, on Nginx 1.0.6

I'm trying to proxy all request to http://www.mydomain.com/fetch to other hosts, the destination URL is specified as a GET parameter named "url".

For instance, when user requests either one:

http://www.mydomain.com/fetch?url=http://another-server.mydomain.com/foo/bar

http://www.mydomain.com/fetch?url=http://another-server/foo/bar

it should be proxyed to

http://another-server.mydomain.com/foo/bar

I'm using the following nginx config and it works fine only if the url paramter contains domain name, like http://another-server.mydomain.com/...; but fails on http://another-server/... on error:

another-server could not be resolved (3: Host not found)

nginx.conf is:

http {
...
# the DNS server
resolver 171.10.129.16;
server {
    listen       80;
    server_name  localhost;
    root /path/to/site/root;

    location = /fetch {            
        proxy_pass $arg_url;
    }
}

Here, I'd like to resolve all URL without domain name as host name in mydomain.com, in /etc/resolv.conf, it's possible to specify default search domain name for the whole Linux system, but it doesn't affect nginx resolver:

search mydomain.com

Is it possible in Nginx? Or alternatively, how to "rewrite" the url parameter so that I can add the domain name?

Server is working perfectly but when I do RDC screen is only black. This happened before and I fixed it by powering off the server via KVM/IP, but I don't want do that again. How can I fix this issue.

OS: Windows Server 2008 R2 Web Edition SP1