Ah, I see!
Perhaps it might be an idea to include the creation of the /tmp directory, too, then? It was only a matter of time before something needed it for initialization (it's required by fuse, hence why I included it in the files.initrd.d/glusterfs.list).
When can I "yum update" to the fixed version? ;)
Thanks.
Gordan
On Fri 16/01/09 10:17 , Marc Grimme <gr...@at...> wrote:
> On Friday 16 January 2009 10:33:06 Gordan Bobic wrote:
> > Err... So how has this been working until now? :-/
> >
> > Marc Grimme wrote:
> > > On Friday 16 January 2009 00:59:17 Gordan Bobic wrote:
> > >> Quick question, embarrasing as it may be - where does /sys
> mount point
> > >> get created in the initrd? I have just rebuilt my initrd and
> there is no
> > >> /sys mount point in it, so mounting /sys fails, which is
> probably why
> > >> udev startup doesn't create the device nodes, and from there on
> it all
> > >> goes wrong. grepping for sys and mkdir comes up with nothing...
> What
> > >> have I broken?
> > >
> > > I think nothing. Yesterday I realised that /sys is not created.
> Just
> > > change boot-scripts/etc/boot-lib.sh function initEnv so that it
> is
> > > automatically created:
> > > echo_local_debug "*****************************"
> > > echo_local -n "Mounting Proc-FS"
> > > is_mounted /proc
> > > if [ $? -ne 0 ]; then
> > > [ -d /proc ] || mkdir /proc
> > > exec_local /bin/mount -t proc proc /proc
> > > return_code
> > > else
> > > passed
> > > fi
> > >
> > > echo_local -n "Mounting Sys-FS"
> > > is_mounted /sys
> > > if [ $? -ne 0 ]; then
> > > [ -d /sys ] || mkdir /sys
> > > exec_local /bin/mount -t sysfs none /sys
> > > return_code
> > > else
> > > passed
> > > fi
> I think there was a dir sys in
> /opt/atix/comoonics-bootimage/boot-scripts/sys.
> That is dir copied into the initrd and mapped to /.
> --
> Gruss / Regards,
> Marc Grimme
> http://www.atix.de/ http://www.open-sharedroot.org/
>
>
---- Msg sent via @Mail - http://atmail.com/
|
