2

I have the networking setup described here:

Why is this static route not taking effect?

Currently, I can ping my DNS server but all DNS requests fail:

~$ ping 130.35.249.52
PING 130.35.249.52 (130.35.249.52) 56(84) bytes of data.
64 bytes from 130.35.249.52: icmp_seq=1 ttl=57 time=105 ms
64 bytes from 130.35.249.52: icmp_seq=2 ttl=57 time=111 ms
^C
--- 130.35.249.52 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1000ms
rtt min/avg/max/mdev = 105.604/108.388/111.172/2.784 ms

~$ nslookup oracle.com 130.35.249.52
;; connection timed out; no servers could be reached

I can think of no good reason for this, because Ubuntu is (to my knowledge) not running a firewall. However, Ubuntu is virtualized, so I hope that it isn't something wrong with the way that its VirtualBox host is set up.

Improve this question
4
  • Can you post the contents of /etc/network/interfaces file in the question?
    – Ramesh
    Apr 4, 2014 at 1:54
  • You should probably add the dns server entry in \etc\resolv.conf file.
    – Ramesh
    Apr 4, 2014 at 1:56
  • /etc/network/interfaces is empty (except for lo); network manager is used. Since I'm explicitly passing the DNS server IP addres to nslookup, I would think that /etc/resolv.conf doesn't matter.
    – Reinderien
    Apr 4, 2014 at 1:58
  • Did you try restarting the network services? Also, you can check if self ping is working.
    – Ramesh
    Apr 4, 2014 at 2:00

1 Answer 1

Reset to default
5

First of all, DNS is primarily a UDP service, not a TCP service. DNS is on UDP port 53; make sure that UDP port is open for incoming connections on the DNS server machine. In addition, DNS can optionally use TCP, which uses TCP port 53, but while DNS can work fine without TCP, it doesn't work without UDP.

Second of all, it's far better to use dig instead of nslookup to debug DNS problems. E.g.: dig @130.35.249.52 oracle.com If you do not have dig, get it with yum install bind-tools (RHEL/Oracle/CentOS 6) or the equivalent command for your Linux distribution.

Indeed, I see your firewall lets port 53 TCP through; dig -t @130.35.249.52 oracle.com works, but dig @130.35.249.52 oracle.com doesn't work because UDP is still blocked.

Improve this answer
5
  • Just curious. If the network is failing, how can yum install bind-tools work?
    – Ramesh
    Apr 4, 2014 at 2:42
  • I know that direct DNS requests to that server work from my host machine. The issue is that somewhere between the guest and the host, the request fails. So I don't think there's a problem between the outer Windows network stack all the way to the server; the problem is somewhere between Linux and VirtualBox.
    – Reinderien
    Apr 4, 2014 at 2:52
  • @Ramesh, only half of the network is failing. The public internet is working correctly on eth0. The corporate intranet on eth1 (and its DNS) is the problem. So yum (actually should use apt-get) works.
    – Reinderien
    Apr 4, 2014 at 2:54
  • VirtualBox is really buggy, especially with any networking for the guest more complicated than "NAT" or "Bridged". It's a lot better to use VMware if possible.
    – samiam
    Apr 6, 2014 at 0:52
  • You're right about UDP. That was the problem - because TCP DNS using dig works just fine. I ended up giving up on the second virtual adapter to my Windows VPN, using network-manager-openconnect instead, and it works just fine.
    – Reinderien
    Apr 7, 2014 at 6:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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