Finished the new Debian Sarge installation.

First boot comes up with Grub loading, please wait....

Error 15

That's where it stops.

Tried changing the BIOS Boot sequence, still the same error.

Any suggestions?
-----------------------------------------------
I booted back up with the Install CD and it allowed me to boot into a "Recovery" mode. During the boot I could see that the SCSI Partitions could not be seen. Something was said about going into a Maintenace mode to fix the problem, but since I do not have a clue as to how to fix the problem with the SCSI Partitions I just hit Continue.

Then it gave me another opportunity to go into a Maintenance Mode and this time I did not see any errors or problems listed. Once again I hit Continue. Debian finished loading and is now installing Apts.

What is "fsck" and how is it used for Boot Maintenance. I read about it in some of the other Posts. Could I use it to fix the problem with the SCSI partitions?

I also read about using some other Boot Loader called lilo?

Thanks

If the system went through the maintainance routine it probably fixed whatever was corrupted. Did you do a hard reboot (press reset or switch the power off) during the install? Do not run fsck on a mounted drive. Normally this would be done on another mounted system or using a boot cd like knoppix. The grub error interpretation of the error is grub was expecting a number but received bad data. Sounds like corruption of the file system and could not find the next stage of the boot process either stage1.5 or stage2. Hopefully it will have sorted the problem out when you finish the install. Ideally use an ext3 file system and not ext2. ext3 uses a journal and makes a record of where all the file pointers are every 5 seconds so a sudden shutdown is usually not too dramatic because it uses the last journal to recover. ext2 is not very forgiving and can cause corruptions which are usually recoverable but not always.

The first reboot was done by Debian, which gave me the error 15.

The second one I accomplished with ctrl+alt+del and reset the BIOS Boot sequence.

The third reboot was done again with ctrl+alt+del and I put the Dedian Sarge CD back in.

The file system was whatever Debian specified, I'm sure it was ext3. However, Debian did not do any automatic maintenance that I could see. When it stopped with an error I was asked if I wanted to do Maintenance or Continue.

The first time it stopped with the "fsck failed" error I tried to do Maintenance and was taken to the /grub command line. Since I didn't have a clue and did not have the manual available I had to do a fourth restart.

That time when it gave me the choice to either do Maintenance or Continue I chose to Continue and received two more error screens with the same choices.

Right now Debian is using APT to install the applications I have chosen.

But, my concern is if GRUB could not see the MBR or partitions for my SCSI drive will these installations be worth anything or will I just have to clean off the hard drives and start again in the morning?

I will not know until I try to reboot the system.

It sounds like there is file system corruption there. If you didn't repair the file system then you will continue to have problems. If you have a copy of knoppix available then use that to do a repair (e2fsck /dev/hdxy) or boot off the cd again and go into maintenance and allow it to repair the file system. You don't say which install you are doing but if its Woody I think the default is ext2. You can change it to ext3 by doing from the command line as root /sbin/tune2fs -j /dev/hdxy where x=drive and y=partition #

The installation was Debian Sarge from the FTP.

When I chose to go into Maintenance Debian just took me to the /grub command line. It did nothing itself, it just sat there waiting for me to type in some command.

I didn't see any attempt by Debian to automatically repair any problem.

Is there a required sequence for creating the different partitions?

I'll search for knoppix tomorrow.

If this is the situation at boot time it may be that grub is not properly configured. You need to have a look at /boot/grub/menu.lst and check that grub is looking at the correct location.

You should have lines like the following;

title Debian Testing GNU/Linux, kernel 2.6.3-1-k7
root (hd2,2)
kernel /boot/vmlinuz-2.6.3-1-k7 root=/dev/hdd3 ro
initrd /boot/initrd.img-2.6.3-1-k7
savedefault
boot

just to explain the configuration;
grub numbers the drives and partitions from 0 so in the line above hd(2,2) it is drive 3 and partition 3 (/dev/hdc3) . You need to check that it is looking at the correct drive and partition and that the vmlinuz line is also correct.

Oh well. Aparently GRUB just will not or can not access the SCSI HDD. It came up with Error 15 again.

If the alternative is something called LILO why doesn't Debian make that available for a choice?

I'll just use my trusty Win98 Boot Disk and fdisk to remove all of the partitions and start again.

I suspect something has gone wrong with the partitioning of the drive. grub is really good and much easier than lilo to use. When you do the next install make sure you use ext3.

I have read a few discussions about the different fs formats available for Linux.

For GRUB I read the best fs is Ext3, which is the default for the Debian Sarge Net Install.

For the rest, I read that Reiserfs is better especially for the larger HDD's we have now.

How can you have mixed File Systems for the same installation? Wouldn't all partitions have to have either Grub or Feiser?

Thanks

No you don't. The mount point in /etc/fstab will specify the fs type. I have multiple partitions via my fstab and have ext3 and two vfat and can mount an ext2 through a manual mount as well. You can only have a single fs type on a single partition but if you had your installation spread over a number of partitions eg /home /usr and /var on separate partitions you can use any fs type on the these partitions provided you specify the fs type in fstab.

Sorry, but I am doing a Debian Sarge Net Install and do not have access to the fstab file until after the installation is complete. For one thing, I understand that Debian uses a RamDisk during the installation and the fstab has not been written yet.

So far, all I have gotten are failed Installs and am attempting to accomplish it one more time.

During the Sarge Install I am offered the option of setting up my Partitions with different File Systems and thought it might work better if I used the Feiser fs for everything but the /boot partition.

When I use the automatic partitioning function Debian Sarge sets all of the partitions up with Ext3 for the fs. I found I do have the option to change them manually.

I just ran into another question I do not see addressed in any of the Forums or the Installation Manual.

Which Partition do I need to set the "bootable" flag on? /boot or just / ? The last time I set it on the /boot partition and that may have been wrong.

I would suggest you approach this as simply as possible. Use a single partition to do the install. I wouldn't complicate things by using multiple partitions at this stage. You make / the boot partition. Where grub installs the boot loader will depend on which drive is attached to the primary disk controller. The grub boot loader should be installed on the first disk otherwise you are going to have to alter the boot order in the bios. Grub should establish where vmlinuz and initrd.img is located for you so the boot loader should be correctly configured. Grub should not be affected by the fs system used as it is simply a boot loader that writes stage1 to the master boot record (mbr). Stage1 will know where to find stage1.5 or 2 whichever it uses. Normally this is located in /boot/grub. Stage2 will initiate the loading of the initrd.img or RAM disk which is located in /boot

Why does the Debian Sarge Installer keep flagging the SCSI HDD as "do not use'???? Which means there is no Bootable HDD, since the IDE is not a bootable drive.

This time I caught Sarge flagging my SCSI as "do not use" before I clicked continue for the Partitioning. I went back and manually corrected the partitions and Debian even reported that the Partitions were formatted and ready to use.

But, during Boot-Up the SCSI drive was found by GRUB and the Partitions were not available!

The SCSI HDD is the only drive set for the system to boot from in BIOS. The IDE is just a data storage drive and is not even listed in the BIOS Boot Sequence.

Again, as I go through the Debian Sarge Install, as soon as I set a partition on the IDE Drive, whether as 1 partition for the whole drive or multiple partitions, the SCSI HDD Partition is changed to "do not use" by Debian Sarge.

WHY??????

Maybe its time for a different flavor of Linux.

I am not sure why this is happening as I have not worked with scsi drives. Because a drive is not set to boot in the bios does not mean the bios or the os cannot see it. I suggest you check the /boot/grub/menu.lst to see what has been configured. Login as root and when you have a prompt do cat /boot/grub/menu.lst | more and have a look at the file.

I'm not sure either.

But, I can't login to look at anything until the installation finishes and boots up. So, I am unable to look at any specific files such as menu.lst because it hasn't been created yet.

This time I installed LILO for the boot loader and the system booted. BUT, now it can not see the IDE HDD again! "FATAL ide_detect not found" message during the Linux loading.

So, I am back to the screen to either hit ctrl+d to exit from the shell or hit enter for maintenance, which means absolutely nothing to me.

If I hit enter I am taken to a command line from which I do not know what Debian expects me to do.

If I hit ctrl+d it will continue booting with erros that are not fixed.

The system has now frozen at "Starting system log daemon: syslogd