On Tue, May 25, 1999 at 03:00:26PM -0400, Steve Cooper wrote:
> 
...
> 
> The reason you don't get a speed up under raid1 is because
> each drive is a physical device. As such there is no advantage
> to asking 1 drive for one chunk of data while asking the other
> for the next chunk of data UNLESS the chunks are large enough
> to offset the time required for the drive to physically reach the next
> piece of desired data.
> 
> Even if the transfer size is 64Kbytes, alternating from drive
> to the other will not yield higher throughput because each
> drive will have the seek to the next 64K chunk. How do they
> reach the next chunk? By passing directly over the 64K chunk
> that is not desired (because its coming from the other drive).
> The time required to skip that 64K chunk is the same as the
> time required to read that 64K chunk.
> 
> However, if you construct a test where there are two independent
> streams of data to be read and the software raid1 implementation
> is written properly, then you should see double the read performance.

Yes, I guess you're right that the way raid-1 stripes the reads doesn't
necessarily yield higher read performance after all...   Here's a little
test I did:

raid-0 on two disks:
              -------Sequential Output-------- ---Sequential Input-- --Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
          900  6160 97.1 21710 73.3  8559 52.5  7841 94.2 23977 63.9 157.3  5.5

raid-1 on the same disks:
              -------Sequential Output-------- ---Sequential Input-- --Random--
              -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine    MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
          470  5801 94.6 11719 39.5  5264 32.5  6931 83.0 11861 34.8 167.4  4.7

Hmm....   I know that raid-1 does distribute the reads to both disks, so I would
think that read-performance should increase.  But it seems like it doesn't. At
least not in this case.   Btw. the disks where on separate SCSI controllers.

................................................................
: [EMAIL PROTECTED]  : And I see the elder races,         :
:.........................: putrid forms of man                :
:   Jakob Østergaard      : See him rise and claim the earth,  :
:        OZ9ABN           : his downfall is at hand.           :
:.........................:............{Konkhra}...............:

Reply via email to