Linux on Highpoint Raid controller??

[doan]

I spent some time trying to get my sytstem to dual boot with winXP and fedora3. I have an Abit BX7-Raid with CD and DVD burners on the two "native" ATA controllers and 2 hardrives on the Highpoint raid controller - not in raid, just 2 drives.

With XP loaded on a 20GB partition of drive 1, I loaded Fedora3 on the the remainder of that 80gb drive. FC3 found and recognized the drives on the highpoint controller and the install completes without any issues other than installing grub took longer than expected.

When I boot into FC3 from the hardrive it works ok but the boot process seems slow. When I reboot or shutdown it posts lots of drive access errors. Booting into XP works fine.

If I move the hard drive to one of the native controllers everything works fine. Since it works at all FC3 must have some drivers for the highpoint controller but I can't find anything about it on the web.

Anyone have experience with linux and the highpoint raid controllers?

Thanks,

David

 

[bersl2]

Perhaps something is assuming that your hard disks reside in /dev/hd[a-d], when they are (probably) /dev/hd[ef] (unless you are using a customized build of the kernel). Or maybe the driver for the RAID card is currently a module and thus is loaded into memory as an initrd at boot, but is unloaded before the kernel completes the final stages of shutting down the system. A large initrd would explain the long periods of time involved with the boot loader and booting in general.

 

[doan]

the Hard drives actually show up as /dev/hde and /dev/hdg (skips f for some reason). I can't check the other stuff since fedora is no longer on this system. (mbr got corrpted and had to rebuild the system to get back to windows)

 

[doan]

Anybody Else?

 

[bersl2]

My recommendation would be to build a kernel with the driver for the controller card build into the kernel and not as a module. I'm thinking that Fedora has the driver compiled as a module and put into an initrd. I don't think this should be causing slowdowns, but it would explain the errors at shutdown if the module is being unloaded before the system has finished putting executables stored on disk into memory.

So either compile the controller's driver into the kernel, or live with these minor annoyances, since it seems as though no damage is being done.

One more question: how long exactly is "slow"? "Slow" compared to XP? We use shell scripts for much of our boot process which, although slower, are much more customizable. Furthermore, while Windows is still loading things well into the presentation of the GUI, giving you the impression that it's fully loaded, most Linux distros makes the user wait until all loading is complete. It's a difference of philosophies (blocking vs. non-blocking, scripted vs. binary).

And I'm guessing that it's not using /dev/hdf because the controller has two IDE channels, and you are not using slaves for either of the channels.