| 1 |
On 12/19/2016 01:09 PM, Andrej Rode wrote: |
| 2 |
> |
| 3 |
>>> It is even more frustrating that these so-called predictable network |
| 4 |
>>> names actually can change on a reboot, it's happened to me more than |
| 5 |
>>> once when multiple network cards are detected in a different order. |
| 6 |
> |
| 7 |
> Then you might found a bug? With predictable network names the name of |
| 8 |
> your device depends on the PCIe slot/address it is in. If you change |
| 9 |
> positions in the board your names should change, not on reboot. |
| 10 |
> |
| 11 |
> And you might believe it or not, running Linux on servers is much more |
| 12 |
> popular than running Linux on your home desktop. Thus I'd guess things |
| 13 |
> tend to be made easier for people with more than one network card. |
| 14 |
> Certainly you don't want rely on random device enumeration order on |
| 15 |
> reboot if you run a webserver with multiple network devices. |
| 16 |
> |
| 17 |
> If you want to disable this on hosts running systemd read [0]. |
| 18 |
> |
| 19 |
> Cheers, |
| 20 |
> Andrej |
| 21 |
> |
| 22 |
> [0] |
| 23 |
> https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ |
| 24 |
> |
| 25 |
|
| 26 |
It could be I found a bug. After a reboot it went from the normal enp0s1 |
| 27 |
(or whatever) to eno1677789 or something ridiculous. I had this happen |
| 28 |
on two different machines. |
| 29 |
|
| 30 |
I tried systemd on a couple machines to see what the hubbub was about, |
| 31 |
and on one machine it crashed on reboot. I probably forgot something, |
| 32 |
but alas, I'd already removed openrc and found out about the binary |
| 33 |
logs. I had no way to download something that supported systemd at the |
| 34 |
time so I could view the log. |
| 35 |
|
| 36 |
Went back to openrc on that machine and disabled that stupid predictable |
| 37 |
name junk. When you only have one network interface, it's rather silly. |
| 38 |
That should have been by default off, server administrators could turn |
| 39 |
it back on. |
| 40 |
|
| 41 |
For what it's worth, systemd works fine on my 11-year-old laptop. It had |
| 42 |
issues with my desktop (which is almost 9 years old now) but I really |
| 43 |
don't know if it was related to mdadm, or ??? I thought I had it figured |
| 44 |
out on my desktop but on reboot systemd wouldn't mark my IMSM mdadm |
| 45 |
array as clean leading to a disk thrashing every time I rebooted, making |
| 46 |
my computer almost unusable. |
| 47 |
|
| 48 |
Dan |
Archives