> Hi,
>
> I have a diskless client with a Fedora Core 5
> server. Everything kind of works on the server:
> PXE-boot, DHCPD, TFTP, etc.
>
> When I use NFS-root as the rootfs for the diskless
> client, then everything works fine. No problems.
>
> But I want to use a ramdisk image gzipped file as
> the rootfs for the client. This seems to work, but
> something strange happens when /etc/rc.d/rc.sysinit
> is processed.
>
> This script hangs in a sort of 'wait' status until
> I type the 'i' key on my keyboard. It then asks for
> each service whether to start or not with a [Y/N]
> question.
>
> Why does the rc.sysinit script suddenly become so
> interactive? Is this a result of the ramdisk rootfs?
> How come?
>
> Thanks,
> Rob.
>
> PS: yet another thing: the rc.sysinit script calls
> /usr/bin/rhgb-client, but at that stage of the
> boot process /usr is not yet mounted on a diskless
> machine !?!?!
Another appendix to my own message above:
When using a ramdisk rootfs on the diskless client,
the rc.d/rc.sysinit script somehow seems to wait
for keyboard input at the end at the line:
/sbin/getkey i && touch /var/run/confirm
For misterious reasons, PROMPT="yes" at this point.
Once again, when I use NFS rootfs instead, this
script does not 'wait' here.
As a kind of remedy, I have commented out the
getkey line for the diskless client rootfs and then
everything goes well.
Apparently, the rc.d/rc.sysinit script overlooks a
certain case when using diskless boot. But what is
that?
The rc.sysinit seems not so 'diskless-boot' friendly.
Regards,
Rob.
____________________________________________________________________________________
Sponsored Link
Get a free Motorola Razr! Today Only!
Choose Cingular, Sprint, Verizon, Alltel, or T-Mobile.
http://www.letstalk.com/inlink.htm?to=592913
