I use Ubuntu 16.04 and connect my machine alongside other machines by using a Hub to a modem to use Internet. Today I encounter with a problem:

Other machine can connect to the Internet; but I cannot access web (Server Not Found error on browser and ping: unknown host google.com as result of $ ping google.com in terminal), unless after doing disconnecting/re-connecting network cable. But this way only works for a few seconds up to 2 minutes. Although my Android device can connect to the web continuously without any problem through WiFi.

Also I connected my machine directly to the modem and even changed the network (Ethernet) cable. But result was the same.

What causes this problem and how I can solve that?


I can open from my machine but cannot login; although it is possible to do that from my other machines.


  • /ect/resolv.conf only contains nameserver (but that file is a link to /run/resolvconf/resolv.conf)
  • the result of $ nmcli device show $ethernet | grep DNS is:
    • IP4.DNS1:
    • IP4.DNS2: -ping works always
  • Both $ dig google.com and $ dig google.com work when I can access web ; but when problem occurs, I get ";; connection timed out; no servers could be reached"


It seems that my machine automatically changes used connection from Internet (which is used to access web) to virbr0 which causes problem:

When I can access web: enter image description here

When problem occurs: enter image description here


Result of $ ip route show when I can access & cannot access the web:

default via dev eno1  proto static  metric 100 dev eno1  scope link  metric 1000 dev eno1  proto kernel  scope link  src  metric 100 dev virbr0  proto kernel  scope link  src linkdown 
  • You can connect to the router but not login, yet you can on other PC's? Curious, I would be looking for a second router, that could explain all of your issues. – acejavelin Aug 3 '18 at 11:02
  • What is the virtual bridge for? – xenoid Aug 3 '18 at 20:04
  • Next step is doing ip route show in both cases. – xenoid Aug 3 '18 at 20:26
  • @xenoid I think virbr0 must be created and used by VirtualBox. – hasanghaforian Aug 4 '18 at 1:40
  • @xenoid Result of $ ip route show in both cases are same. I added result to my question. – hasanghaforian Aug 4 '18 at 1:54

Likely a DNS problem..

  • check that your /ect/resolv.conf only contains nameserver,
  • issue nmcli device show $ethernet | grep DNS to list the nameservers ($ethernet is the name of your interface, normally some enp*)
  • check if you can access these (ping...)
  • check if they can be used dig @dnserver somesite.com

The connection editor lets you to force a DNS in your connection, for instance:

  • (Google)
  • (OpenDNS)
  • Thank you for your reply. I updated my question, please see it again. Also it seems you are sure that the problem is not in about hardwares, for example network card. Did I guess true? – hasanghaforian Aug 3 '18 at 10:34
  • Since it works when you disconnect/reconnect then no too likely tthe hardware. Does disconnecting/reconnecting Ethernet using the systray applet also fixes (temporarily) the problem? Otherwise, look in /var/log/syslog around the time you loose the DNS resolution. Also, does it work if you connect to a totally different network (or not using the hub) ? Is everything in DHCP without any possibilty that two machines get the same IP address? – xenoid Aug 3 '18 at 11:56
  • I checked my connection and /var/log/syslog. Now I edited my question please see it again. – hasanghaforian Aug 3 '18 at 18:37

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.