Red Hat Bugzilla – Bug 5250
9-9-1999 boot.img and bootnet.img break IBM 7000 Adaptec support
Last modified: 2015-01-07 18:38:10 EST
Original RedHat 6.0 boot floppies for install recognized the
Adaptec SCSI controller in IBM Netfinity 7000 M10 system.
Updated 9-9-1999 boot.img recognizes the Adaptec controller
during boot but fails to recognize it when locating disks
for the installation process.
This needs to be 'unbroken'.
What's the problem with the original images that you need
the updated ones for?
We are planning to use IBM ServerRAID as soon as it is supported, are
currently using the built in Adaptec SCSI instead. I can work with
the old images for now, but when the ServerRAID support comes out of
beta it would be nice if the install floppy images that worked with it
still worked with the Adaptec too.
Aside from that I'm philosophically opposed to newer software working
worse than it's predecessor and thought someone at RedHat might care
OK, I was just checking to make sure that the boot disk
wasn't breaking on something it was explicitly designed to fix.
We haven't noticed problems with the Netfinities here;
what are the exact error messages? Does the bus scan just
reveal nothing? Does it detect a different number of
Adaptec controllers (if there are multiple ones)?
Current configuration has only the Adaptec 2940 SCSI adapter
installed. The one that comes standard with IBM Netfinity 7000 M10.
Boot process on 9-9-99 boot.img follows:
AIC7895 SCSI BIOS 1.34 at boot finds the drives as IBM-PSG 9103LC
Welcome Splash page - OK
select Intel EtherExpress Pro 100
network config details set
enter NFS server details
** A screen flashes for a fraction of a second "scanning...2940"
dialog box "SCSI Configuration"
Do you have any SCSI adapters?
Default response is no, I select yes
select Adaptec 2740,2840,1940
Dialog box "ERROR"
I can't find the device anywhere on your system
Loops back to SCSI Configuration dialog
You need to boot with 'linux supp' on the boot disk, or update
the modules on your NFS install image with ones from the
updated supplementary image.
Tried 'linux supp' at install lilo boot prompt but with identical
results except for requiring 2 more floppies during boot.
Hmm; I can't reproduce this here on our 7000; booting with
'linux mem=256m supp' (the mem argument is only needed if
you have a LOT of memory, and it would die much sooner if
you needed it), and then using the supp.img & modules.img
from the errata, it detects the 2940 and detects the two
drives attached to it.
What are the error messages on VC's 3 and 4?
Closing bug due to lack of activity.