RAID-3? (was: cvs commit: src MAINTAINERS)

[Pawel Jakub Dawidek]

On Thu, Aug 19, 2004 at 12:13:59PM +0930, Greg 'groggy' Lehey wrote:
+> On Tuesday, 17 August 2004 at 15:16:12 +0200, Pawel Jakub Dawidek wrote:
+> > On Tue, Aug 17, 2004 at 10:10:20PM +0930, Greg 'groggy' Lehey wrote:
+> > +> On the contrary.  RAID-3 requires byte-level striping, which is
+> > +> ridiculous on the hardware that FreeBSD supports.
[...]
+> > Want to compare performance with vinum's RAID5?:)
+> 
+> Feel free.  But do it with more than a single process accessing the
+> disks.

Tests were done using this HW:

CPU: Pentium II/Pentium II Xeon/Celeron (534.55-MHz 686-class CPU)
  Origin = "GenuineIntel"  Id = 0x665  Stepping = 5
  Features=0x183fbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR>
real memory  = 67108864 (64 MB)
avail memory = 60256256 (57 MB)
MPTable: <OEM00000 PROD00000000>
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
[...]
iir0: <ICP Disk Array Controller> mem 0xd8100000-0xd8103fff irq 17 at device 13.0 on pci0
[...]
da0 at iir0 bus 2 target 0 lun 0
da0: <ICP Host Drive   #00 > Fixed Direct Access SCSI-2 device 
da0: Tagged Queueing Enabled
da0: 8675MB (17767890 512 byte sectors: 255H 63S/T 1106C)
da1 at iir0 bus 2 target 1 lun 0
da1: <ICP Host Drive   #01 > Fixed Direct Access SCSI-2 device 
da1: Tagged Queueing Enabled
da1: 8675MB (17767890 512 byte sectors: 255H 63S/T 1106C)
da2 at iir0 bus 2 target 2 lun 0
da2: <ICP Host Drive   #02 > Fixed Direct Access SCSI-2 device 
da2: Tagged Queueing Enabled
da2: 8675MB (17767890 512 byte sectors: 255H 63S/T 1106C)

The software you can find under /usr/src/tools/tools/raidtest/ in HEAD
branch.
Test has been done for 10000 random requests (offset, size and operation
type was random).

RAID3:
		Number of	Bytes per	Requests per
Operations	processes	second		second
----------------------------------------------------------------------
READ		3		6329500		95
READ		15		8981047		135
READ		100		10719314	161
WRITE		3		5073263		76
WRITE		15		7467387		112
WRITE		100		8631136		129
READ/WRITE	3		6041795		90
READ/WRITE	15		8104847		121
READ/WRITE	100		9494250		142

RAID5:
		Number of	Bytes per	Requests per
Operations	processes	second		second
----------------------------------------------------------------------
READ		3		6041795		90
READ		15		14768833	222
READ		100		19546985	294
WRITE		3		3479568		52
WRITE		15		5403231		81
WRITE		100		6211191		93
READ/WRITE	3		4521071		68
READ/WRITE	15		7911875		119
READ/WRITE	100		9360528		140

As you can see RAID5 is only faster in READ tests for 15 and 100 processes
working in parallel. As I can see, RAID3 is faster in all the rest tests.

Of course you are welcome to try by yourself.

Anyway, if I can ask for something. Think twice before calling something
ridiculous without understanding.

PS. I wonder about read optimization, so parity component can be also
    used for reading in round-robin fashion...

-- 
Pawel Jakub Dawidek                       http://www.FreeBSD.org
pjd at FreeBSD.org                           http://garage.freebsd.pl
FreeBSD committer                         Am I Evil? Yes, I Am!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-performance/attachments/20040820/648b2f50/attachment.bin