At 02:05 PM 12/27/2005, Michael Stone wrote:
On Tue, Dec 27, 2005 at 11:50:16AM -0500, Ron wrote:
Sorry. A decade+ RWE in production with RAID 5 using controllers
as bad as Adaptec and as good as Mylex, Chaparral, LSI Logic
(including their Engino stuff), and Xyratex under 5 different OS's
(Sun, Linux, M$, DEC, HP) on each of Oracle, SQL Server, DB2,
mySQL, and pg shows that RAID 5 writes are slower than RAID 5 reads
What does that have to do with anything? That wasn't the question...
Your quoted position is "there isn't a 'RAID 5 penalty' assuming
you've got a reasonably fast controller and you're doing large
sequential writes (or have enough cache that random writes can be
batched as large sequential writes)."
My experience across a wide range of HW, OSs, DBMS, and applications
says you are wrong. Given enough IO, RAID 5 takes a bigger
performance hit for writes than RAID 10 does.
Enough IO, sequential or otherwise, will result in a situation where
a RAID 10 array using the same number of HDs (and therefore of ~1/2
the usable capacity) will have better write performance than the
equivalent RAID 5 built using the same number of HDs.
There is a 'RAID 5 write penalty'.
Said RAID 10 array will also be more robust than a RAID 5 built using
the same number of HDs.
OTOH, that does not make RAID 5 "bad". Nor are statements like
"Never use RAID 5!" realistic or reasonable.
Also, performance is not the only or even most important reason for
choosing RAID 10 or RAID 50 over RAID 5. Robustness considerations
can be more important than performance ones.
cheers,
Ron
