1 |
Neil Bothwick <neil@××××××××××.uk> writes: |
2 |
|
3 |
> On Wed, 21 Dec 2016 22:48:29 +0100, lee wrote: |
4 |
> |
5 |
>> > You can't switch any two names because the udev rules are run singly, |
6 |
>> > so at one point you will be trying to rename an interface with a name |
7 |
>> > that is already in use. |
8 |
>> |
9 |
>> I mean more like renaming them on the fly --- or by having a |
10 |
>> configuration file with key:value pairs like 'enp69s0f1:eth3' --- or |
11 |
>> perhaps triples like 'enp69s0f1:eth3:"DMZ Interface"'. |
12 |
> |
13 |
> In that case you may as well leave the unique names in place and set up |
14 |
> recognisable aliases. |
15 |
|
16 |
Sure, you can call the names you pick aliases. Can that be done? Not |
17 |
as in "going back to the old way", but as described. |
18 |
|
19 |
>> That way, you could have a recognisable name (or several names) for |
20 |
>> every unrecognisable one and assume that "eth3" or "foo" or however you |
21 |
>> want to call it is the same interface just as much as you would with |
22 |
>> unrecognisable names --- plus the advantage that when you ever need to |
23 |
>> change an interface, you only need to edit one small file rather than |
24 |
>> various configurations files having the unrecognisable name(s) in them. |
25 |
> |
26 |
> There are no config files to edit with the predictable names, the |
27 |
> names are created from the physical location of the port. That's why |
28 |
> they are called predictable, |
29 |
|
30 |
I only know what the names are when I can look them up when the computer |
31 |
is running. I don't call that "predictable". |
32 |
|
33 |
They were much more predictable before because I could be reasonably |
34 |
sure that each of the ports would be called 'ethN', starting with N = 0, |
35 |
unless I changed a card for a different one after an udev rule had |
36 |
already been created. Now I can only assume that they will be called |
37 |
something. |
38 |
|
39 |
> unless you move the NIC to a different PCI slot, it will always have |
40 |
> the same name, no matter what other hardware you add or remove. Yes, |
41 |
> the names are cumbersome, but they have to be like that to guarantee |
42 |
> their uniqueness. |
43 |
|
44 |
You don't need to defend the unrecognisable names. The names used for |
45 |
referring to network ports don't need to be like that. |
46 |
|
47 |
The perceived advantage lies in being able to refer to network ports in |
48 |
a more reliable way, and I don't see how using unrecognisable names |
49 |
instead of recognisable ones would make anything easier. |
50 |
|
51 |
It would have made things easier if the problem had been solved by |
52 |
giving them recognisable names (or aliases) by default --- or even if |
53 |
the default names (aliases) were the same as the unrecognisable names |
54 |
--- and allowing to easily configure the names (aliases) actually used |
55 |
to refer to the ports. |
56 |
|
57 |
Being able to refer to things in more reliable ways improves the quality |
58 |
of the software. Using unrecognisable names for things reduces the |
59 |
quality. |
60 |
|
61 |
This is like you're defending a type of new pliers. The old ones didn't |
62 |
hold stuff as securely as the new ones do, but the new ones require that |
63 |
you use both hands to use them. The new pliers can provide an advantage |
64 |
for instances in which you do have to hold something very securely --- |
65 |
and in which another tool, like a vice, might be more appropriate anyway |
66 |
--- but for most of the time, they hinder you doing your work because |
67 |
they're so unwieldy. Of course, you call the new pliers "more secure |
68 |
pliers" rather than "unwieldy pliers", because that makes them easier to |
69 |
sell. |
70 |
|
71 |
Alas, "improvements" just like this seem to become more and more common, |
72 |
replacing actual improvements: The king gets new garments not seldom |
73 |
times, yet twice a day, and those who cry deceit are called not children |
74 |
but trolls. |
75 |
|
76 |
But who knows, perhaps it is now possible to easily, on the fly, name |
77 |
the network ports through a neat configuration file. I'm merely asking |
78 |
if there is because I don't know and would find that very useful. |
79 |
|
80 |
> How often you you have to type interface names anyway, and how many of |
81 |
> those are in a shell with tab completion that takes care of it for |
82 |
> you? |
83 |
|
84 |
None of them are, and I don't type the names. They require copy and |
85 |
paste, or very careful and tedious typing after looking them up. |