Dell/Perc5 raid/MPT SAS Integrated Raid Write Performance

Sat May 19 01:49:00 UTC 2007

I am using a LSI-1064E
2 boards, each with firmware images I compiled from a base firmware
image concatenated with an NVDATA file that you configure yourself.
I set up the ports as narrow and attach 2 SATA drives in RAID-1
configuration and I get 28-30 MBytes/sec.
I compare this to the same drives running on an SATA capable IDE
controller and the performance is the same.
So RAID-1 does not seem to impact it ... if the volume is in sync.
I have the sync rate set to about 2% default in the firmware config
(NVDATA) file, and when the volume is out of sync, the performance drop
to about 15 MBytes/second and jump back t full speed once in sync.

Attaching SAS 2.5 inch drives I get 55-60 MBytes/sec.

Using firmware revision 01.18.00.00 with MPI Version 1.5.13.0 using
NVDATA file version 25.
and on the other board I am experimenting with a beta phase9 (whatever
that means) firmware, compiled with MPI version 1.5.14.0 and NVDATA file
version 28.
On both setups I removed the LSI BIOS part (mptsas.rom) after a while,
since I do not need it.
I configure the controller from the OS itself. This all with Matthew's
driver ported to an FreeBSD 2.1.0 variant, including the NEW CAM layer.

The firmware for 01.18.00.00 is freely available from their site.  
http://www.lsi.com/storage_home/products_home/standard_product_ics/sas_i
cs/lsisas1064e/index.html

Use the host bus adapter firmware if you want to try it out:
http://www.lsi.com/cm/License.do?url=/files/support/ssp/fusionmpt/sas/bi
os_firmware/SAS3041E-R_IR_10-30-06.zip&prodName=LSISAS3041E-R&subType=Fi
rmware&locale=EN

Johan

-----Original Message-----
From: owner-freebsd-scsi at freebsd.org
[mailto:owner-freebsd-scsi at freebsd.org] 
Sent: Friday, May 18, 2007 7:11 PM
To: freebsd-scsi at freebsd.org
Cc: freebsd-performance at freebsd.org
Subject: Dell/Perc5 raid/MPT SAS Integrated Raid Write Performance

A lot of users have seen some very poor write performance on the MPT 
(LSI-Logic) driven SAS/SATA controllers, particularly those that 
function in Integrated Raid (mirroring) mode.

Some of the reported performance issues are pretty clearly single 
spindle small transfer IOPS issues. For example, directory intensive and

small file operations like mail server applications can do very poorly 
on single spindle SATA drives that are connected via a SAS channel that 
doesn't enable write cacheing on the SATA drive (i.e., does not flow 
through WCE for SCSI emulation, as the LSI-Logic *apparently* does not).

Benchmarks like Postmark show pretty amazing differences when run on a 
PATA or native SATA based drive (1000s of ops/second) and on single 
Fibre Channel or SCSI drives (100s of ops/second) and can be even worse 
for SATA drives on a SAS controller.

In these cases, there isn't much to be done- the h/w being picked 
doesn't match the application.

However, other users have reported things which are *clearly* bad 
performance issues. In these cases users have reported sequential write 
speed to be a small fraction of read speed. That is, a single threaded 
read of a 10GB file will get spindle rotational speed magnitude for the 
disk in question (~40MB/s) but will only write at around ~6MB/s. This is

clearly broken and wrong.

Since I don't actually have a *lot* of MPT h/w and none that shows this 
write performance problem could folks do me a favor and at the next 
reboot get into the LSI-Logic BIOS utilities and find me all the 
firmware revision numbers? This might help me nail down some differences

to go talk to LSI-Logic about.

The overriding LSI-Logic BIOS revision is of interest, but also any of 
the firmware revision numbers. For example, the loaned Sunfire 4100 I 
have has 6.0.2.0 for the BIOS revision, but 1.04.00.00-IR for the 
firmware- and this system, which has two integral SAS 2.5" drives, 
writes at 50MB/s with them set up as an integrated mirror.

-matt

_______________________________________________
freebsd-scsi at freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-scsi
To unsubscribe, send any mail to "freebsd-scsi-unsubscribe at freebsd.org"