Recover from a virtual Ethernet network failure

Learn how to recover from a virtual Ethernet network failure after a kernel upgrade.

If you have upgraded to a 64-bit kernel (from a 32-bit kernel less than 2.4.10) or to a 32-bit kernel level greater that 2.4.10, your network device information may have changed. The notation of the virtual network devices in Linux has changed from vethXY to ethXY.

To understand what network devices relates to a respective virtual Ethernet line description, you can view the display message log using this command:

dmesg | fgrep veth | less

The output of this command should produce messages similar to the following:

veth.c: Found an Ethernet device eth0 (veth=0) (addr=c000000000ff2800)

In this case, the message is telling you that the i5/OS® virtual line description veth0 now relates to a Linux net device of eth0, and veth=0 correlates to VLAN0 in i5/OS. If for some reason your display message log is overrun, you can also analyze the network devices in the proc fileserver with the following command:

cat /proc/iSeries/veth/[netdevice]

Running this command with a valid netdevice will produce output similar to this:

Net device: c000000000ff2800
Net device name: eth0
Address: 0201FF00FF01
Promiscuous: 0
All multicast: 0
Number multicast: 0

This file is telling you that the Linux net device eth0 maps to the first virtual LAN on the i5/OS which is VLAN0.