1 |
On Thu, Jan 5, 2012 at 5:08 PM, Nicolas Sebrecht |
2 |
<nicolas.s-dev@×××××××.net> wrote: |
3 |
> On Thu, Jan 05, 2012 at 02:20:21PM -0500, Michael Mol wrote: |
4 |
> |
5 |
>> FWIW, I had a /dev/cdrom symlink long before *devfs* even existed, let |
6 |
>> alone udev. |
7 |
> |
8 |
> We are not looking for device paths that existed berfore udev. Actually, |
9 |
> most of them exist since much more time than udev. It's not relevant at |
10 |
> all. |
11 |
|
12 |
You missed my point. My point was that udev wasn't needed to resolve |
13 |
the use case you described, that stable solutions to such cases |
14 |
preceded udev, and so udev wasn't a necessary tool to solve them. |
15 |
|
16 |
> |
17 |
>> Also, ethN numberings are generally stable until and unless you do |
18 |
>> some strange BIOS tweaking or hardware changes, and should be able to |
19 |
>> be stabilized in the event the instability comes from some racy module |
20 |
>> loading mechanism. |
21 |
> |
22 |
> This is not true. I've had computers in hands where network cards could |
23 |
> change of names without any BIOS tunning. |
24 |
|
25 |
Did this happen after a kernel update or a change to your kernel |
26 |
configuration? A software update? A change in the set of enabled |
27 |
modules, or which were built-in vs built as modules? |
28 |
|
29 |
In any production server environment, I would assume you're already |
30 |
watching the thing like a hawk and verifying that the thing comes up |
31 |
properly after a reboot. Reboots should be very, very rare things. I |
32 |
try to do things more or less correctly on a high-profile machine, and |
33 |
I'm still giddy when the once-in-a-blue-moon reboot doesn't break |
34 |
anything. |
35 |
|
36 |
> BIOS is a executed program and |
37 |
> the way each is implemented can guarantee *or not* to have the |
38 |
> conditions for persistent NIC names on Linux. |
39 |
|
40 |
What you're saying is that NIC stability is dependent on how the OEM |
41 |
built the BIOS and software. I'll posit there may be strange NICs out |
42 |
there which can't be initialized within a deterministic time frame |
43 |
without some external factor such as a link handshake. If this were a |
44 |
common behavior for a piece of hardware though, I'd consider it |
45 |
indicative of shoddy quality, and would want to replace it. |
46 |
Regardless, it's resolvable in software without using a tool that |
47 |
imposes significant restrictions on system structure. |
48 |
|
49 |
> |
50 |
>> udev's attempts at stabilizing network interfaces have made things |
51 |
>> worse more often than I've heard of it making them better. Hit any |
52 |
>> search engine for "eth0 missing 70-persistent-net.rules". |
53 |
> |
54 |
> It's fully expected and required. Persistent naming can work if you have |
55 |
> a configuration for that somewhere. I see nothing worse here. |
56 |
|
57 |
One week, I helped no fewer than five people who ran afoul of the |
58 |
70-persistent-net.rules file, and didn't know why their eth0 |
59 |
disappeared. These weren't newbie Linux users, either. Some knew their |
60 |
way around GNOME better than I still do, and they mostly knew their |
61 |
way around the shell. Some were networking professionals pulling more |
62 |
than I do. |
63 |
|
64 |
I'd wager the vast majority of systems out there have devices named as |
65 |
'ethN' for wired connections, and 'wlanN', 'athN' or whatever for |
66 |
wireless connections. And that the vast majority of those systems have |
67 |
one or fewer wired connection ports. And that the further vast |
68 |
majority of those don't have customized 70-persistent-net.rules files |
69 |
as you and I have. If that's true, then the persistent-net rules |
70 |
behavior currently harms more users than it benefits. |
71 |
|
72 |
> But I see |
73 |
> an improvement to let me tune the NIC names if I need to. I have routers |
74 |
> with *lot of* NIC cards where this feature is very usefull (expressive |
75 |
> names are much better than ethX). |
76 |
|
77 |
I, too, noted this as a potential advantage of udev. On my router, I |
78 |
have five interfaces. 'wan', 'he-tunnel', lan, wifi, lo and 'tun0'. |
79 |
tun0 is only so-named because it's an OpenVPN thing I haven't bothered |
80 |
to change. I've tried to advocate use this feature of udev. |
81 |
|
82 |
But I administer my router the way I like to. Most people I've pointed |
83 |
toward this capability just go "Meh. I have a list of interfaces and |
84 |
what they're for." even when they already have udev. |
85 |
|
86 |
-- |
87 |
:wq |