On Sun, 2002-10-06 at 13:06, Frank Tanner III wrote:
> Unfortunately, all of my Linux boxes here at gome are
> now borked because of the bum burn.
Frank,
Not sure how that happened, unless you intentionally tried to install
8.0 on them from known questionable CDs. Dogged consistency is not
always A Good Thing. <g>
Be that as it may, do you have access to *any* previous RH releases
(production or beta) v7.2 or later? If so, do a minimal install on one
of your systems and give it a static IP address in your network. It is
not necessary to burn CDs or install X to follow these 10 simple steps:
1. Make sure you have cdrecord and NFS installed.
2. Download fresh ISOs of CDs 1-3 from your favorite mirror:
a. $ cd
b. $ mkdir psyche
c. $ cd psyche
d. $ ftp <your.mirror.of.choice>
logon as user 'ftp' or 'anonymous'
enter your e-mail address as the password
ftp> hash <- display # progress marks
ftp> bell <- beep when transfer complete
ftp> prompt <- turn off prompting
ftp> binary <- this is critical!
ftp> cd <iso.directory>
ftp> get MD5SUM
ftp> mget psyche*.iso
very long wait here...
if interrupted, repeat 2d. thru cd <iso.directory> steps
ftp> reget <full.name.of.interrupted.iso>
another long wait...
ftp> get <any.remaining.iso.file(s)>
With T1 access and a fast mirror, it will take about six hours
to download all three. With 28.8 kbps dial-up access (like me),
you'll be downloading iso files non-stop for several days.
3. Confirm the integrity of the iso files you've downloaded
$ md5sum -c MD5SUM
4. As root, mount disc1 to access the floppy images
# mount /home/<you>/psyche-i386-disc1.iso /mnt -t iso9660 -o loop
# cd /mnt/images
5. Create a network installation boot floppy
# dd if=bootnet.img of=/dev/fd0 bs=2048
and maybe supplemental block and NIC driver floppies
# dd if=drvblock.img of=/dev/fd0 bs=2048
# dd if=drvnet.img of=/dev/fd0 bs=2048
6. Make /home/<you>/psyche available as an NFS export
# vi /etc/exports
Add this line:
/home/<you>/psyche *(ro)
Confirm it's been added
# exportfs
Look for
/pub/<you>/psyche <world>
7. Reinitialize NFS
# service nfs restart
8. Boot one of your "borked" machines with the network boot floppy.
If you need a special HDD or NIC driver, hit F2 when asked and insert
the supplemental block mode and/or NIC driver floppy.
9. If you already have a DHCP server on your network, use it. Otherwise
pick an unused IP address in your network.
10. Select the NFS install option, and when asked, enter the following
Name of NFS server: NFS.server.IP.addr
Directory: /home/<you>/psyche/ <- be sure to use
a trailing slash!
anaconda should start...
If you follow these steps, I think I can just about guarantee a
successful Red Hat 8.0 installation on each of your "borked" home
machines without having to burn a single CD.
One really nice thing about having a 100 Mbps Fast Ethernet switch is
that you can do about a dozen NFS installs at the same time without
slowing things down very much. This is because RPM database operations
on destination machines are usually a lot slower than Fast Ethernet
transfers from the source. Red Hat uses NFS installs for their RHCE
classroom machines.
Hope this helps...
--Doc Savage
Fairview Heights, IL
