How to boot after changing IDE connector
 

Hello,

due to better cable management I had to change the connection of my IDE-HDD-drives on the mainboard from IDE1 to IDE2.

Since then booting fails (not surprising). I receive:

On the machine, I have installed Knoppmyth R5F27 (on Debian GNU 4.0/Etch)

Is there a way to fix this, using some sort of live distribution?

Thank you and kind regards.

Rai

On newish systems you could change the boot disk order so that ide channel 1 is now above channel 0. The issue is that now this second or third in order drive is trying to use the older relative paths.

I'd see if bios can change the order then you have only that to fix. Otherwise you have to boot to a live cd and fix all the issues with paths.

In BIOS the correct HD with LILO is identified so booting starts at least.

But for my understanding, as soon as it comes to the point, where paths are indicated, it will fail (old path hdaX (primary master) is now hdcX (secondary master))

I am thinking of editing /etc/fstab. Some more settings to fix?

[edit:]
Meanwhile I tried to boot LILO using some command line option - but without success. What I did:

1. interrupt LILO by hitting TAB during boot.
2. in the command line I added some options, which I identified from /etc/lilo.conf. So the line looked like this:
But still no success. Same error message as mentioned in the beginning

Desperately, I changed the IDE connectors back, so the HDD is attached to first IDE and it is identified again as /dev/hda.

Now everything works again...
[/edit]

Hi NightStranger

Yep the above advice is correct, the system may boot from the nominated drive (BIOS) but the installed software (and referenced data) will have absolute references to the drives (hd<a><i>), and they would have to be identified and changed manually.

If you go down that road, then I suggest that you boot the system with a Live CD/DVD linux, and then perform global searches for references, but that will not cover any software applications that have been built (compiled) containing absolute references (picked up from system data during a build, or manually specified).

In short, my opinion is that you are always better off in the longer term (many considerations) if you store your valuable user data on a separate physical drive (yes I know that isn't always possible, but more often that not it is :) ), so that if you perform a change of physical drives, or drive order, you avoid this situation.

In this case I would recommend backing up your valuable user data, and performing a fresh OS and application software install, then restore your backed up user data. The other advantage of this approach is that you will have (a) a complete backup copy of your data, and (b) proven that you can recover/rebuild your Linux installation. Both steps of which would be required if you (heaven forbid) suffered a system of hard disk failure) in the future.

Hope that assists.

C.

Member Response
 

Hi,

It does sound as if the OP should be using cable select(cs) for the drive identification on a 40 style cable. Some drives allow 'single/master' thus the placement on the cable doesn't matter in a single drive system. I prefer the 80 conductor for better transfer & configuration. No reflection or signal injection from adjacent conductors when using the 80 style cable.

Norm is to place the master drive at the end of the cable for master/slave configurations or single drive installs. HD IDE Drive jumper configuration should provide some insight.