0

I have a fresh install of Server 14.04 running on an old Dell OptiPlex 755 MT. It automatically connects to my network and I can reach it through SSH without an issue. I also made sure to edit the /etc/network/interfaces to add the auto eth0 and iface eth0 inet dhcp, which I will set a static IP once I'm ready to place the server.

I run

sudo apt-get update && sudo apt-get upgrade

and install the 100+ updates available and reboot. When it comes back up, eth0 is no longer connected and I have to run

sudo ifconfig eth0 up

and dhclient to bring it back up each time. The /etc/network/interfaces file still shows my eth0 information, but the OS doesn't seem to read it.

Is there a certain update from the initial set that is causing this? If so, which one is it so I can remove it? Apologies ahead of time, I'm fairly new to Linux and haven't been able to track down a solution.

4
  • Your system may be using Consistent Network Device Naming, which names network interfaces determined by how they're attached to your motherboard. Were there any other interfaces listed in /etc/network/interfaces (other that lo and virbr0)?
    – Arronical
    Jul 13, 2015 at 14:08
  • May we see: ifconfig and also: cat /etc/network/interfaces Welcome to askubuntu.
    – chili555
    Jul 13, 2015 at 14:39
  • All I have in ifconfig is the lo after applying updates. In /etc/network/interfaces I have 'auto lo' 'iface lo inet loopback' followed by 'auto eth0' 'iface eth0 inet dhcp'. I did a fresh install on another machine without doing the apt-get update/upgrade and it connects with no issue. It's only after I apply the initial updates I lose connectivity and have to manually start the interface and run dhclient.
    – masker671
    Jul 14, 2015 at 14:17
  • UPDATE: I added a wireless card to test a fresh install. It appeared in the /etc/network/interfaces and I applied a static address to it and it connected fine. I downloaded and applied the first round of patches and the same thing happened. Both eth0 and wlan0 do not load and have to be manually started after every reboot after the patches are applied.
    – masker671
    Jul 17, 2015 at 19:03

1 Answer 1

0

Had the same problem. Fresh install on a HP microserver gen8. Works like a charm, network was called em1. In the log it showed that udev had renamed eth0 to em1.

Afther update with aptitude, the rename was gone from the log files, and I had an eth0 interface.

Sollution one:
sudo vim /etc/udev/rules.d/70-persistent-net.rules

change the line:
SUBSYSTEM=="net", ACTION=="add", ...... KERNEL=="eth*", NAME="eth0"

into
SUBSYSTEM=="net", ACTION=="add", ...... KERNEL=="eth*", NAME="em1"

Sollution two
You can also change /etc/network/interfaces.

Modify the two lines that read em1 into eth0.

Don't know what is best. I was used to eth0 from my previus Slackware server so I sticked to eth0.

Do think it is bad that a LTS build changes network names on a update!! Using the ILO evaluation license I can now access the console, but otherwise I would have had to attach a keyboard an screen to fix this.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .