> Do you have DevFS compiled in AND set to mount itself on boot? Both > are necessary for the kernel to see that device.
Yes, and Yes. Again, reconfirmed by redoing the compile. > Barring that, try the old way of referring to it. According to the > kernel docs, it's path would be "/dev/cciss/c0d0p7" if you had the /dev > properly setup. Or try "/dev/hda7", maybe it shows up as a normal IDE > drive to the kernel too. I'll give this a try. Red Hat definitely refers to the device as c0d0p7 rather than disc0/part7. Gentoo didn't seem to. I'll try it though. Ok, tried hda7. that's a no-go. Same error message. (except /hda7 rather then cciss/disc0/part7) Ok, tried c0d0p7. Same as above. I will say, if I try to prompt for anything after the ...boot=/dev/ I just get errors. I assume it's because the partition isn't mounted, but it actually sounds like boot= is the problem, more than the path to the device... Could I be screwing up the boot= part of the equation? If I leave it out completely, then I also get the above message, though I sort of expect it. I should also maybe mention that I also get an error, kmod failed to exec /sbin/modprobe -s -k scsi_hostadapter, error 2 a bit earlier in the process. I googled it, and basically found that this was irrelevant noise, and should be ignored. The thing that worries me though is that most of the times I read about that, people were using IDE devices. I'm actually using SCSI. Also, I chrooted back into the installed version, and added cpqarray cciss to the modules.autoload file. It shouldn't matter though, since it's not at that point when it pukes, and they shouldn't be necessary, since they're compiled into the kernel (smart2 adapter). Kev.
