On Sat, 6 Mar 1999, Adam D. McKenna wrote:
> From: Seth Leigh <[EMAIL PROTECTED]>
>
> :Also, with "raid" having originally stood for "redundant array of
> :inexpensive disks", I find it ironic that people are paranoid about using
> :ide disks for a raid, and instead stick with expensive scsi disks. Thus
> :they have a "RAED". For RAID 5, which can survive the loss of one disk
> :without a hitch, and with the ability to have a hot spare which can be
> :patched into the RAID automatically should one of the "hot" drives fail,
> :I'd say that the much less expensive UDMA drives would work just fine.
>
> "at the risk of starting a religious argument..."
>
> IDE raid is worthless due to the inability of IDE to read from and write to
> multiple drives at the same time.
Maybe in the past, but since UDMA host adapters are using modified SCSI
chipsets to service EIDE drives, your argument is weakening slowly until
EIDE matches SCSI.
Ultra66 is as ugly as SCSI (hardware aspect) with an 80 conductor ribbon
and built in terminating resistor packs to cancel ringing and bounce.
This sounds very SCSI-like, yes-no?
> http://www.adaptec.com/technology/benchmark/scsivudma.html
>
> If you just want to protect against loss of data, then IDE raid is fine. But
> not if you want to get real performance out of your array.
You are a little behind on all the latest IDE changes, but you probably
don't have need to keep up, and that is okay.
There is a new widget from Promise that is called "SuperTraks".
It is and I2O bus that is a RAID 0,1,4,5 with hot-swap that is potentially
OS independent as a hardware-raid solution. I have known about this for
about 3 month prior to the public release, while trying to get Promise to
work with me directly for support under Linux for their "FastTrak".
Cease-Fire..............
Andre Hedrick
The Linux IDE guy -- http://www.dyer.vanderbilt.edu/server/udma/
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
