On Wednesday 19 Dec 2001 00:05, Oden Eriksson wrote:
> On Wednesdayen den 19 December 2001 00.26, Guillaume Cottenceau wrote:
> > Oden Eriksson <[EMAIL PROTECTED]> writes:
> > > On Tuesdayen den 18 December 2001 15.21, Chmouel Boudjnah wrote:
> > > > Oden Eriksson <[EMAIL PROTECTED]> writes:
> > > > > What if I had /usr/local as... ntfs (silly I know), that fs
> > > > > module won't be in the ramdisk.
> > > >
> > > > so what, after loading the / partitions it will load the others
> > > > modules for others partitions.
> > >
> > > Yes, I understand that. The problems I have had occurs when I
> > > forget to fix lilo...
> >
> > Oden -- but what is the problem exactly ?
> >
> > What's in ramdisk is only for mounting the / partition, no more
> > is needed, so if you're using ext3, reiserfs module is not needed
> > in ramdisk.
>
> A number of times when I have upgraded the kernel, I forgot to fix the
> depmod and the ramdisk stuff manually. In these rare cases it leaved me
> with a kernel without modules, other than those in the ramdisk (no nic,
> etc.). I must add that the automization doing rpm "-U" or "-i" of the
> kernel works much better now. As I use my md for backups and other
> important stuff I have to have it accessible at all times. Since my
> recent change to ext3 on "/", the fact that my md is at "/mnt/md0" and
> rfs, I now need to make the ramdisk manually anyway.
>
> What I'm trying to say is... if one uses things like md, it would be
> nice to have mkinitrd check what fs the md is using, and include (or
> propose to include) that module in the ramdisk, regardless of its
> mountpoint.
>
> Do I make any sense at all?

It might make more sense to me if I knew what "md" was.

-- 
Peter Ruskin, Wrexham, Wales.
Registered Linux User No. 219434 ( see http://counter.li.org/ ).
Mandrake Linux release 8.1 (Vitamin) for i586
Kernel 2.4.8-34.1mdk-win4lin,  XFree86 4.1.0, patch level 21mdk.
KDE: 2.2.2.  Qt: 2.3.2.  Uptime 0 hours 46 minutes.
------------------------------------------------------------------

Reply via email to