[mythtv-users] ide controller advice
Jarod Wilson
jarod at wilsonet.com
Sun Dec 25 18:12:53 EST 2005
On Sunday 25 December 2005 14:09, John Andersen wrote:
> On 12/16/05, Jon Whitear <jon at whitear.org> wrote:
> > My research indicated that 'cheap' IDE RAID controllers aren't worth it
> > - they do the RAID in software (the driver) anyway, so you might as well
> > use Linux software RAID. If you want RAID in hardware, the 3ware
> > controllers are well supported.
>
> Exactly. Software Raid is stable and scales with the processor
> when you upgrade
Software RAID takes up hardly any cpu at all, even w/RAID 5. Scaling with the
processor isn't something to worry about if you have any cpu made in the last
three or four years. The only big cpu hits are if you have to reconstruct an
array.
> whereas the hardware raid is as fast on day
> one as it will ever get.
That's not exactly true. Some hardware RAID controllers perform better on one
motherboard or another for various reasons. For example, I have a PCI-X LSI
MegaRAID 320-1 controller that is absolutely pathetic (max 15MB/s througput)
on a dual Xeon motherboard, but performs about 5x better on a dual Opteron
motherboard (same drives in both cases).
> Still, it sounded like the OP was not using anything other
> than raid 0 (concatenation) and this is just as well done with
> LVM as with any form of raid.
RAID 0 is NOT concatenation, it is striping. A concatenated array means data
is first written to drive 1, once its full, writing begins on drive 2, and so
on. A striped array means data is written to all drives in the array at once,
in parallel ("striped" across the disks).
But yes, cheap SATA/IDE RAID controllers are mostly worthless.
--
Jarod Wilson
jarod at wilsonet.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mythtv.org/pipermail/mythtv-users/attachments/20051225/3307d5e7/attachment.pgp
More information about the mythtv-users
mailing list