[nycbug-talk] Quasi-RAID [Was: AMD Raid question. . .]

[Charles Sprickman]

On Mon, 24 Oct 2005 alex at pilosoft.com wrote:

> Hint: Most of the "hardware" raid is *not* hardware raid.
>
> http://linux.yyz.us/sata/faq-sata-raid.html
>
> Controller fired.

I'd like to fork this debate off so George can go about his business while 
I ask tangential questions...

I've never understood the "quasi" hardware RAID that's popping up on more 
and more mainboards.  It does not fit either of the traditional 
definitions:

-Hardware RAID - you have either an add-on card, a ZCR card (Zero 
Channel RAID - leverages built-in scsi controllers), or an outboard 
scsi-scsi/ide-ide box.  All of these have a processor and drive 
controllers onboard to handle the RAID calculations and logic.  They hide 
the raw drives from the OS and present the configured RAID array instead.

-Software RAID - You use standard IDE/SCSI controllers and the OS deals 
with creating/maintaining the array.  The host CPU is used for all the 
XOR calculations when running RAID 5.  Booting may or may not be simple 
when a drive fails as you have to rely on the BIOS to boot off of the 
"good" drive if an element of the array fails.  This seems less complex 
with non-i386 processors as they don't have a clunky, hacked BIOS to deal 
with.

-Quasi-RAID - What George apparently has here and many people now call 
"Hardware RAID".  How does this work?  Is it basically a standard IDE 
controller?  Do they just load up a special BIOS that has enough smarts to 
know what to boot off of when there's a failure?  Why is the array 
"configured" in the BIOS?  How are the drives presented to the OS?

I'd love to know more about this, as it seems like it may be appropriate 
(and cheap) in some situations...

Thanks,

Charles

> -alex
>
> _______________________________________________
> % NYC*BUG talk mailing list
> http://lists.nycbug.org/mailman/listinfo/talk
> %Be sure to check out our Jobs and NYCBUG-announce lists
> %We meet the first Wednesday of the month
>