Redhat 8 Installation

[jamesb]

So langsam wird das Forum richtig bekannt... Ich habe vorhin eine Email bekommen, von einem Spanier ;)

Da ich ihm nicht weiterhelfen kann, poste ich die Mail einfach mal hier (ich hoffe, hier hat niemand was gegen Englisch):

Anyway, I have a rootserver at Schlund but have a lot of problems installing Redhat.
I've been trying to install Redhat using the texts I found on rootforum.org/ (http://www.rootforum.org/forum/viewtopic.php?t=3461) and for Redhat 8.0 installation a text by Heiko Schulz (http://www.myrootserver.de/redhat/redhat-install.txt).

Specially the last text has been very helpfull but not helpfull enough to get Redhat 8.0 installed.

After installing everything my server just doesn't reboot in normal bootmode. If I then boot in rescue mode to check boot.log there isn't even a boot.log in /var/log.

Here you see the whole installation process I use:

# booting in rescue mode
mke2fs -j /dev/sda1
mke2fs -j /dev/sda3
mkswap /dev/sda2

mount /dev/sda3 /mnt
mkdir -p /mnt/boot /mnt/etc/ /mnt/tmp/install /mnt/var/lib/rpm
mount /dev/sda1 /mnt/boot

echo > /mnt/etc/fstab "/dev/sda3 / ext3 defaults 0 1"
echo >> /mnt/etc/fstab "/dev/sda2 none swap sw 0 0"
echo >> /mnt/etc/fstab "/dev/sda1 /boot ext3 defaults 1 2"
echo >> /mnt/etc/fstab "proc /proc proc defaults 0 0"
echo >> /mnt/etc/fstab ""

echo > /mnt/etc/lilo.conf "boot=/dev/sda"
echo >> /mnt/etc/lilo.conf "root=/dev/sda3"
echo >> /mnt/etc/lilo.conf "image=/boot/vmlinuz"
echo >> /mnt/etc/lilo.conf "label=RedHat"

cd /mnt/tmp/install
wget http://myrootserver.de/redhat/install.log
grep installiert.$ install.log | cut -f1 -d' ' > file.list
echo >> file.list "openssh-3.4p1-2"
echo >> file.list "openssh-server-3.4p1-2"
echo >> file.list "ncftp-3.1.3-6"
echo >> file.list "wget-1.8.2-3"
echo >> file.list "netconfig-0.8.12-3"
wget ftp://ftp.rpm.org/pub/rpm/dist/rpm-4.0. ... 386.tar.gz
tar -C / --exclude *.html --exclude *.h -xzf rpm-4.0.2.i386.tar.gz
awk '{printf "if ! wget ftp://ftp.cs.tu-berlin.de/pub/linux/Mir ... s.i386.rpm; then wget ftp://ftp.cs.tu-berlin.de/pub/linux/Mir ... noarch.rpm; fi;nrpm --excludedocs --root /mnt -ivh %s* --nodepsn", $0, $0, $0}' file.list > install
echo >> install "wget ftp://ftp.cs.tu-berlin.de/pub/linux/Mir ... 4.i586.rpm";
echo >> install "rpm --root /mnt -ivh kernel-2.4.18-14.i586.rpm"
chmod 700 install
./install

chroot /mnt

echo >> /etc/modules.conf "alias eth0 8139too"
echo >> /etc/modules.conf ""

echo >> /etc/sysconfig/network "NETWORKING=yes"
echo >> /etc/sysconfig/network "HOSTNAME=s1234567.rootmaster.info"
echo >> /etc/sysconfig/network "IP=1.2.3.4"
netconfig

passwd

lilo

rm -rf /tmp/install
exit
umount /dev/sda1
cd /
umount /dev/sda3

# setting bootmode back to normal
reboot


After that I wait for 20 minutes or so and try to ping my server, nothing...
Then I reboot in rescue mode and do:

mount /dev/sda3 /mnt
mount /dev/sda1 /mnt/boot
chroot /mnt
ls /var/log

and I see there isn't even a boot.log!

What am I doing wrong???

I hope you can help because I'm out of ideas.

Soweit die Email.

JamesB

[fritz]

Code: Select all

echo >> /etc/sysconfig/network "IP=1.2.3.4" 

is that a dummy IP ?

regards Fritz

[captaincrunch]

You've got SCSI-Disks in your Server, which can't be found, because you'll have either :
- to load the correct scsi-module in modules.conf
- build an initrd (but I don't know, how you could do this in the rescue-system or
- install a Kernel with built-in Support (not as a module) for your SCSI-Controller (where I don't know, what Controller is installed)

[silly]

Hallo zusammen :)

na JamesB bist du jetzt dran ;)

You've got SCSI-Disks in your Server, which can't be found, because you'll have either :
- to load the correct scsi-module in modules.conf
- build an initrd (but I don't know, how you could do this in the rescue-system or
- install a Kernel with built-in Support (not as a module) for your SCSI-Controller (where I don't know, what Controller is installed)

das habe ich ihm auch schon versucht klarzumachen mit den modules aber irgendwie kommt er da net ganz mit :( ....

[misterff]

Hello to everybody!

I wanted to start with a "Thank you" to Silly for answering my emails (however I don't recall you mentioning anything about modules for scsi devices, but you really did help on the way), and to JamesB for posting my email here. I'm glad english is no problem.

Fritz: Yes that's a dummy IP.

CaptainCrunch:
After trying different kernels I called Schlund to ask which kind of SCSI-controller my rootserver uses, they tell me I just got IDE hd's! They said that the 'sda' is part of the RAID system or something like that.

Can somebody explain that to me? And how do I have to adapt the installation of Redhat to that??

[captaincrunch]

Hi misterff !

I just looked this up, and the hotline-guy was (partly) right : Schlund Rootservers have built-in IDE-Drives, not SCSI-Devices. The drives are configured as a RAID-1, where I don't know, wether they use Hard- or Software-Raid.

If they use Hardware-RAID (what I'm thinking), you just have to change the /dev/sda's to /dev/hda, then everything should be fine. If not, just tell us and we'll see what we can do.

Btw. : Sorry for my bad english ...

[misterff]

I already tried that.

While trying:
mke2fs -j /dev/hda1

it says:
mke2fs: Device not configured while trying to determine filesystem size

What I also tried was to change only sda to hda in lilo.conf and fstab but at the end, when you start lilo it gives an error too..

Could you also explain to me what this RAID is and what the difference is between hardware and software RAID?

[captaincrunch]

Sorry, I habe any experience with (IDE-) RAID-Controllers and the way they are handled by the kernel.

I think it would be the fastest solution if you let the guys at Schlund & Partner reinstall your server. Afterwards you could exactly see the configuration, which makes reinstalling much more easier.

Could you also explain to me what this RAID is and what the difference is between hardware and software RAID?

OK, I'll have a try :
RAID stands for Redundant Array of Inexpensive Disks, which means, that you put together two or more disks to a disk-group. With this disk-group you have several possibilities. At Schlund & Partner you have RAID-Level 1 (Mirror) :
Two or more disks are "mirrored". You see both disks as one and all data written to them gehts written to both disks at the same time, so your data won't be lost if one disks fails.

With Hardware-RAID this task is done by a piece of hardware, which is faster, because it handles everything by itself.
Software-RAID ist handled by the Kernel and of course uses system-ressources for this.

You can find more about the hole topic here : http://www.tldp.org/HOWTO/Software-RAID-HOWTO.html

[misterff]

Thank you for the explanation CaptainCrunch.

I just called Schlund. Reinstalling the server would be no use. I would have the same problem. They use always the same images and I would get the whole damn Ensim stuff on my server again. Which was the reason for me to reinstall everything (because they said they couldn't install Redhat without Ensim).

By the way, they use hardware RAID-1 at Schlund. I don't have a clue why that makes it so damn hard to get a basic Redhat 8.0 on my server. Maybe you guys have some tips for me. Meanwhile I'll take a look at this software-raid HOWTO. Goddamn...

[silly]

I just called Schlund. Reinstalling the server would be no use. I would have the same problem. They use always the same images and I would get the whole damn Ensim stuff on my server again.

I think CaptainCrunch meant you should do the reinstall to get more information about your System, so its easier to install a diffrent ;)

[misterff]

Oh, I understand.

Well, I saved a lot of the files of the initial installation. Specially from the /etc directory (not from /boot though). Here's what lilo.conf looks like:

Code: Select all

boot=/dev/sda
map=/boot/map
install=/boot/boot.b
prompt
timeout=100
vga=normal
default=linux

image=/boot/vmlinuz
        root=/dev/sda3
        label=linux
        read-only

By the way, I just found a lilo.conf.hda too! Interesting... Looks like this:

Code: Select all

boot=/dev/hda
map=/boot/map
install=/boot/boot.b
prompt
timeout=100
vga=normal
default=linux

image=/boot/vmlinuz
        root=/dev/hda3
        label=linux
        read-only

And I found the file fstab but also fstab.hda with the same kind of differences in it like lilo.conf and lilo.conf.hda. (Just hda instead of sda)

Anybody familiar with this?

[captaincrunch]

I'd say now, that /dev/sda should be right because I think that the IDE-RAIDcontroller ist handled by the kernel as an SCSI-Device and because lilo.conf is the default setting ... also, you already told us that mkfs wouldn't work on hda

The easiest way would be to let Schlund reinstall your server, afterwards you could look up the configuration (lspci -vv | mail you@domain.es for example). When you got everything together it will be no problem to install RedHat 8 ...

[misterff]

Schlund reinstalled the server for me.

I looked at what lspci -v gives, but I'm not sure where to look to be certain I will be able to reinstall redhat 8. I looked at /var/log/boot.log but that doesn't help me much either and it's not that I can just try different things because I would have to call Schlund after every try! Can somebody help me on the way?

Here's what lspci gives me (in non-verbose mode):

Code: Select all

00:00.0 Host bridge: VIA Technologies, Inc. VT8601 [Apollo ProMedia] (rev 05)
00:01.0 PCI bridge: VIA Technologies, Inc. VT8601 [Apollo ProMedia AGP]
00:07.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super South] (rev 40)
00:07.1 IDE interface: VIA Technologies, Inc. Bus Master IDE (rev 06)
00:07.4 Bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 40)
00:09.0 RAID bus controller: 3ware Inc: Unknown device 1001 (rev 01)
00:0d.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139 (rev 10)
01:00.0 VGA compatible controller: Trident Microsystems CyberBlade/i1 (rev 6a)

[bükkes]

Conquering the same situation, I also tried various kernels and included the 3ware-module into the initrd - without success. ;-)

So, did you finally solve the problem?

kind regards,
Thorsten

[Hoffe, Posting in Englisch ist ok...]

[bükkes]

A second 'rpm -ivf kernel-xxxx' in chrooted mode works, obviously lilo has to be run first.

HTH,
Thorsten