Random failure to resolve internet names

Support
,
#1

NethServer Version: 7.6
Module: network / dns
Hi all,
I’m running neth on a PC with 2 NICs as a gateway.
One NIC is configured as red and is connected to my ISP supplied HFC router / modem. I have configured this NIC to receive an IP from the router’s dhcp service (192.168.0.x).
The other NIC is configured as green (10.0.0.1) and the internal network DHCP is setup on this one.
I have not created any bridges as I don’t really need AD for the network here.
The named servers are configured to use 8.8.8.8 and 8.8.4.4.

Since I have installed Nethserver, I have had intermittent issues resolving internet server names (I can still ping external servers using ip addresses, but their domain names).
Typically, this happens at least once a day and seems to happen when the network is under load (I have 3 teenagers…….)
Any support would be appreciated

#2

Hi Keith,

please check (with nslookup on command line or with diagnostics panel in Nethserver):

  • if the clients use the right DNS (10.0.0.1 in your case)
  • if the Nethserver can resolve the names correctly.

You may try to configure a static ip on the NIC just to exclude that the modems dhcp overrides the DNS servers Nethserver is using.

#3

Hi Markus,
Thank you for your reply.
While troubleshooting, I had set the red IP address to static and reserved the IP on the router.
The nethserver as well as all of the clients experience the inability to resolve names though I can ping external addresses,.
Nslookup fails to resolve the domain names when I have attempted it from both the command line and through the web UI.
When the issues is being experienced, the external IP address displayed in the diagnostics page is also lost.

#4

Does it work if you lookup directly?

nslookup nethserver.org 8.8.8.8

Does it help if you restart the modem? Are there errors in /var/log/messages?

You may turn on dnsmasq query logging:

Edit /usr/lib/systemd/system/dnsmasq.service and add the -q option in line 6:

ExecStart=/usr/sbin/dnsmasq -k -q

Apply the config:

systemctl daemon-reload
systemctl restart dnsmasq

Check the log or use “Log viewer” in web UI:

grep dnsmasq /var/log/messages

Don’t forget to remove it again after testing to avoid having a lot of messages…

This is the command executed when fetching the external ip address:
dig +short myip.opendns.com @resolver1.opendns.com

1 Like
#5

Hi Marrkus,
Thank you for suggesting the lower level logging.
I have discovered what was happening. The log files indicated that the outgoing request was attempting to resolve using the green interface. After some investigations, the issue seems to be the driver that was being used on the red NIC. I swapped out the NIC and reset the network configuration and it looks like we are back up and running.

Thanks for your help