Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: Performance Question: RAID or Individual drives ?
>Art S. Kagel wrote in message <35AB7297.3954_at_bloomberg.net>...
>You are right that a stripe set with large stripe size (64K) is bad for
>random reads. However, tests have shown that for smaller stripe sizes
>(say 16K) random performance is not as bad. (BTW 16K happens also to
>be the favorite I/O write request size for Informix, and I believe for
>Oracle also so this is a better match to the server's activity anyway.)
Yes and no. Smaller stripe size can negate the effect of striping problems on randon I/O, but on AIX the Oracle recommended stripe size is 64K, and that seems to be the figure mentioned for most Oracle/UNIX installations. Latest Oracle datawarehousing book recommends much, much larger stripe sizes - my testing has shown this to actually make things worse on AIX. Just shows the value of testing! :-)
>In addition, keep in mind we are talking about a disk farm for a
>relational database not for a filesystem. Oracle and Informix have
>impressive buffer caches and do significant and intelligent read-ahead
>so that their disk accesses are rarely random at all! Even writes are
>coallesced so that writes are in large sequential blocks.
True. But of course there is always a cut-off. My stuff is datawarehouse, large reads and large monthly updates, so striping is fine for me. Other's should approach it with caution, although it *may* too be the right solution for them.
>Sorry but
>for Oracle and Informix RAID10 is just clearly superior, more expensive
>for sure but better. And for installations where cost is a factor
>RAID3 is superior to RAID5 for both performance and safety. Let the
>LVM do the RAID10 or RAID3 it is worth the administration overhead.
RAID0+1 is a very nice solution. I didn't mention this in my latest testing as the adapter (SSA) doesn't support it. Only problem is it [RAID0+1] is very, very expensive for a 1TB database! Still, for a smaller system (or a larger budget) it's the way to go.
>
>> >It looks to me like Motox's testing is using a single Wide or Fast only
>> >controller and is hitting his controller's I/O limit (10MB/s).
>>
>> Nope, SSA. 4 x 20MB/s ports, 2 loops, per adapter.
>
>Sounds cool, but it does not explain your results, I see a bottleneck
>somewhere. Is the RAID set limited to a single controller or spread?
>Just curious.
>
>Art S. Kagel
Yes, it's an interesting one. I find that the drives will never run at anything like peak when striping, whether in software (LVM) or hardware, RAID5. I suppose I'll have to chase IBM up on this one, as NT Disk Manager striping over SCSI was much better.
Cheers,
MotoX. Received on Tue Jul 14 1998 - 11:36:53 CDT
![]() |
![]() |