Web lists-archives.com

Re: Naming of network devices

 ❦ 10 juillet 2017 15:53 -0400, Marvin Renich <mrvn@xxxxxxxxxx> :

>> > With the new scheme, if I want to rename the interface to something more
>> > meaningful, I have to go find an older machine that already has a
>> > persistent-net.rules file or read through a lot of documentation to
>> > figure out the correct syntax.  I then have to determine the correct
>> > ATTR elements to identify the interface in question, and type all of
>> > that information by hand, hoping I type everything correctly.
>> Have a look at systemd.link(5) which enables you to do that without
>> debugging udev.
> Okay, I had a look on a six-weeks-before-release stretch VM, and I don't
> see anything there that makes this easier than with the udev .rules
> file.  There is still no skeleton .link file that already has the
> appropriate attributes already filled in.  It looks like I still have to
> create a .link file and manually type the appropriate attributes.  Am I
> missing something?

Not really.

In /etc/systemd/network/whatever.link, put:



Note the manual page has a similar example. I don't see exactly how this
is more complex than the udev rules we had:

SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*",
ATTR{address}=="00:11:22:33:44:55", ATTR{type}=="1", KERNEL=="eth*",

> BTW, there seems to be a typo in the man page:  it refers to a
> /run/systemd/network directory; should that be /run/systemd/netif/links?
> I'll leave you to file a bug if appropriate.

No, it seems correct. All systemd-related stuff look at /lib/systemd/X
(shipped with a package), /run/systemd/X (added by a third-party
program) and /etc/systemd/X (added by the user).

>> > There is an easy fix to revert the default behavior while still allowing
>> > knowledgeable sysadmins to get the new behavior.  On the other hand,
>> > those who need to administer systems but are not sysadmins by trade (and
>> > thus will have to do significantly more research to even know that the
>> > older behavior is possible) are the ones who need the older behavior as
>> > the default.
>> Having a default behavior that's unreliable doesn't seem great
>> either. The change is surprising (and not expected on non-new systems),
>> but there is nothing difficult in identifying the right interface with
>> the new naming scheme. Other major distributions are using this new
>> scheme (notably Ubuntu which has no reason to have users smarter than
>> ours) and I don't see a lot of drama on their side. It seems that for
>> some reason, we are the only ones making a fuss about any change.
> I don't get this at all.  The previous behavior was neither unreliable
> nor unpredictable (unless you are talking about much older systems
> before persistent-net.rules).

It was unreliable. Google for "eth0_rename", you'll get ton of examples
of people with hosts that don't boot reliably because of the inherent
race conditions. Udev people didn't invent all this just to get people
pissed. They have fixed a real problem.

> What change is surprising?  The change between jessie and stretch?
> Absolutely.


> You have completely sidestepped the question, which is about the
> relative importance of the two goals, simple names _all the time_ vs
> avoiding a state file.  The previous behavior sacrifices the second,
> much less important goal in favor of the first.  The new behavior
> sacrifices the first in favor of the second.  Until you address this
> issue, all your explanations look like attempts to ameliorate bad
> behavior.

Predictable names are more important to me. Simple name should also work
for most people (using the eno*) scheme. Maybe it's not available as
widely as it should be.
Make sure input cannot violate the limits of the program.
            - The Elements of Programming Style (Kernighan & Plauger)

Attachment: signature.asc
Description: PGP signature