Bug in Initscripts...?

semi linux linuxsemi at gmail.com
Fri Aug 10 23:51:06 UTC 2007


On 8/6/07, Bill Nottingham <notting at redhat.com> wrote:
> semi linux (linuxsemi at gmail.com) said:
> > I'm not sure if this is a "bug", something that has been left-out or
> > an intentional design decision.  So feel free to tell me if I'm out to
> > lunch.
> >
> > In FC4, initscripts-8.11.1 will not rename an interfaces based on
> > HWADDR unless ONBOOT=yes...
> >
> > I'm upgrading to F7 (or RHEL 5) soon but I was wondering the current
> > status of this issue.  I scoured the Changelog on RHEL5
> > (initscripts-8.45.14.EL) and got nothing.  Before I pull down the CVS
> > history and follow the code path I was wondering if anyone had dealt
> > with this before.
>
> It's very much intentional, as with udev interfaces can be loaded
> in a random order. So renaming them all on module load is simplest.
>
> Bill

"on module load"?  I assume you mean I'm supposed to use udev rules?
Not all modules have a naming param, do they?

If it's udev rules, I've been using lines like this:
KERNEL=="eth*", SYSFS{address}=="00:90:fb:xx:xx:xx", NAME="eth50"
(very annoying note: they are case sensitive!)

Some daemon (kudzu?) tries to remember which port was which so this
naming/renaming thing is far from clear unless you know of an ironclad
way of doing it...  I thought initscripts would be the magic bullet,
but I guess not.

- G.




More information about the fedora-devel-list mailing list