On 11/6/16 11:15 AM, Rowland Penny wrote:

> One of my servers crashed because of a motherboard problem, but, as
> luck had it, there was something on the HD I was working on and I
> hadn't fully backed up.
>
> I stuck another motherboard in and started up the machine again, up it
> came, after fsck'ing the HD and everything worked, apart from the
> network. Checked lspci etc and as far as I could see, there was nothing
> wrong, but I just couldn't get eth0 to work (did I say there was only 1
> network card?)
>
> Finally, in desperation, I ran 'dmesg | grep eth0' and found my problem:
>

> root@server:~# dmesg | grep eth0
> [    0.921998] r8169 0000:02:00.0 eth0: RTL8168b/8111b at 0xffffc90000006000, 00:1d:60:fc:29:e6, XID 18000000 IRQ 41
> [    0.922001] r8169 0000:02:00.0 eth0: jumbo features [frames: 4080 bytes, tx checksumming: ko]
> [    7.620169] systemd-udevd[362]: renamed network interface eth0 to eth1

>
> Why, oh why, did systemd-udevd rename eth0 to eth1 ????????
>
>

As others have pointed out, this behavior dates back to udev. (And boy
did that mess me up the first time it happened.)

For me, the first warning signs about systemd was that the perpetrator
of udev was one of the key perpetrators of systemd.

Miles Fidelman


--
In theory, there is no difference between theory and practice.
In practice, there is. .... Yogi Berra