I'm running Debian sarge with a custom-built 2.6.15 kernel. I have an array of 10 73G SCSI drives connected to an Adaptec 29160 card. According to the output of dmesg, all 10 disks are detected correctly, but something about Debian means that only 6 of them get entries in /dev. I get /dev/sda - /dev/sdf, but nothing more. I think I overcame this once on an Ubuntu box, but I can't remember how or what causes it. Anyone have any ideas? Thanks in advance.

Well, what about /sys/block - how many sd* do you see? Do you mind editing /etc/udev/rules.d/... to manually check that everything is taken into account or to even view the dev-files in /sys/block/sd? and manually create devices?

Thanks for the reply. I'm behind the times on things like udev and /sys - I only had the vaguest of ideas as to what they were for. I don't have the drives connected right now, but I'll be reconnecting them soon, and then I'll take a look in /sys. I actually don't know if udev is even installed - I looked at another Debian sarge minimal install I have here and it's not installed by default. I'll post here again if I still can't figure this out. Thanks again.

A follow-up:

I finally got around to experimenting some more with this. Turns out my problem was my own cluelessness - Debian sarge doesn't install udev by default! Its /dev directory is simply full of manually-created device nodes. When I used mknod to create the appropriate devices, everything was fine.

So now I'm running software RAID with an Adaptec 29160. It works, but once a drive fails it seems that the rebuild of the hot spare sucks all the bus bandwidth and causes the array to be spectacularly unresponsive (45 seconds - 1.5 minutes for an ls, for example). But that's another problem altogether...