Tests of RAID adapters

[Francois Tigeot]

On Mon, Apr 25, 2011 at 07:38:57PM +0200, Sascha Wildner wrote:
> On 4/25/2011 9:12, Francois Tigeot wrote:
> >LSI SAS 3081E-R
> >---------------
> >
> >The RAID1 volume created in the BIOS of the card was visible but there were
> >some timeout error messages from the start:
> >
> >I couldn't install DragonFly, newfs_hammer hung at 90% completion.
> 
> I've put an update for mpt(4) here:
> http://leaf.dragonflybsd.org/~swildner/0001-mpt-4-Sync-with-FreeBSD.patch

Thanks for updating the driver. With this new version, things are much
better :)

I've used a similar setup as before, with the following changes:
- host system is running DragonFly 2.10 + updated mps driver
- host system is a 4-core Xeon E5506
- the drives are 2x WD5003ABYX (500 GB, 7200 RPM, 64MB cache)

Observations:

1. no timeouts on startup

mpt0: <LSILogic SAS/SATA Adapter> port 0xe000-0xe0ff mem
0xfbbf0000-0xfbbfffff,0xfbbec000-0xfbbeffff irq 16 at device 0.0 on pci3
mpt0: MPI Version=1.5.20.0
mpt0: Capabilities: ( RAID-0 RAID-1E RAID-1 )
mpt0: 1 Active Volume (2 Max)
mpt0: 2 Hidden Drive Members (14 Max)

mpt0:vol0(mpt0:0:0): Settings ( Hot-Plug-Spares )
mpt0:vol0(mpt0:0:0): Using Spare Pool: 0
mpt0:vol0(mpt0:0:0): 2 Members:
      (mpt0:1:2:0): Primary Online
      (mpt0:1:1:0): Secondary Online
mpt0:vol0(mpt0:0:0): RAID-1 - Optimal
mpt0:vol0(mpt0:0:0): Status ( Enabled )
(mpt0:vol0:1): Physical (mpt0:0:1:0), Pass-thru (mpt0:1:0:0)
(mpt0:vol0:1): Online
(mpt0:vol0:0): Physical (mpt0:0:2:0), Pass-thru (mpt0:1:1:0)
(mpt0:vol0:0): Online

da0 at mpt0 bus 0 target 0 lun 0
da0: <LSILOGIC Logical Volume 3000> Fixed Direct Access SCSI-2
device 
da0: 300.000MB/s transfers
da0: 476837MB (976562176 512 byte sectors: 255H 63S/T 60788C)
no B_DEVMAGIC (bootdev=0)


2. newfs_hammer completes succesfully, although performance is bad

$ iostat da0 2
      tty             da0             cpu
 tin tout  KB/t tps   MB/s  us ni sy in id
   1  245  0.00    0  0.00   0  0  2  0 98
   0   40 16.00  118  1.85   0  0  0  0 100
   0  643 16.00  116  1.81   0  0  0  0 100


3. blogbench also runs without crashing

$ blogbench -d /mnt/bench

[...]
Final score for writes:          1407
Final score for reads :        244892

$ iostat da0 2
      tty             da0             cpu
 tin tout  KB/t tps   MB/s  us ni sy in id
   0   22 15.99 3045 47.55  11  0 89  0  0
   0   11 15.86  865 13.39  11  0 89  0  0
   0   11 15.98 1886 29.44  11  0 89  0  0
   0   23 16.00 2974 46.46  10  0 90  0  1


For information, this is the blogbench score of a disk of the same model
plugged on one of the AHCI SATA ports:

Final score for writes:          1307
Final score for reads :        215165

-- 
Francois Tigeot