ProLiant Servers (ML,DL,SL)
cancel
Showing results for 
Search instead for 
Did you mean: 

Smart Array 6i - very poor read performance

 
GregLS
Occasional Visitor

Smart Array 6i - very poor read performance

I've got a DL360 G4p server with onboard Smart Array 6i with BBWC and two 146GB 15K drives in a mirrored setup. This server is showing very poor performance to the RAID array, at only 40MB/sec. I was expecting something at least in the neighborhood of 80MB/sec - 100MB/sec for the array... Is this normal? Need some ideas. Have already done full firmware update.

OS: CentOS 5.5 x86-64

# hdparm -t /dev/cciss/c0d0p2

/dev/cciss/c0d0p2:
Timing buffered disk reads: 126 MB in 3.02 seconds = 41.76 MB/sec

ACU "config detail":

Smart Array 6i in Slot 0 (Embedded)
Bus Interface: PCI
Slot: 0
Cache Serial Number: P9A3A0C9SY12KU
RAID 6 (ADG) Status: Disabled
Controller Status: OK
Chassis Slot:
Hardware Revision: Rev B
Firmware Version: 2.84
Rebuild Priority: Low
Expand Priority: Low
Surface Scan Delay: 15 secs
Surface Scan Mode: Idle
Post Prompt Timeout: 0 secs
Cache Board Present: True
Cache Status: OK
Accelerator Ratio: 50% Read / 50% Write
Total Cache Size: 192 MB
No-Battery Write Cache: Disabled
Cache Backup Power Source: Batteries
Battery/Capacitor Count: 1
Battery/Capacitor Status: OK
SATA NCQ Supported: False

Array: A
Interface Type: Parallel SCSI
Unused Space: 0 MB
Status: OK

Logical Drive: 1
Size: 136.7 GB
Fault Tolerance: RAID 1
Heads: 255
Sectors Per Track: 32
Cylinders: 35139
Strip Size: 128 KB
Status: OK
Array Accelerator: Enabled
Unique Identifier: 600508B1001FFFFFA01399FD6D6D0002
Disk Name: /dev/cciss/c0d0
Mount Points: /boot 502 MB
OS Status: LOCKED
Logical Drive Label: A01399FD6D6D
Mirror Group 0:
physicaldrive 1:0 (port 1:id 0 , Parallel SCSI, 146.8 GB, OK)
Mirror Group 1:
physicaldrive 1:1 (port 1:id 1 , Parallel SCSI, 146.8 GB, OK)

physicaldrive 1:0
SCSI Bus: 1
SCSI ID: 0
Status: OK
Drive Type: Data Drive
Interface Type: Parallel SCSI
Transfer Mode: Ultra 320 Wide
Size: 146.8 GB
Transfer Speed: 320 MB/Sec
Rotational Speed: 15000
Firmware Revision: HPB5
Serial Number: 3KN0X9CS00007623VC3S
Model: COMPAQ BF1468A4CC

physicaldrive 1:1
SCSI Bus: 1
SCSI ID: 1
Status: OK
Drive Type: Data Drive
Interface Type: Parallel SCSI
Transfer Mode: Ultra 320 Wide
Size: 146.8 GB
Transfer Speed: 320 MB/Sec
Rotational Speed: 15000
Firmware Revision: HPB1
Serial Number: DN01P5C006GW0549
Model: COMPAQ BF14689BC5

5 REPLIES
Michael A. McKenney
Respected Contributor

Re: Smart Array 6i - very poor read performance

No-Battery Write Cache: Disabled

Check the battery and enable the write cache.



GregLS
Occasional Visitor

Re: Smart Array 6i - very poor read performance

> No-Battery Write Cache: Disabled
>
> Check the battery and enable the write cache.

As far as I know the "No-Battery Write Cache" option controls whether the 6i enables the write cache if the battery is missing or fails. It doesn't mean the write cache is disabled or the battery is missing.

Battery/Capacitor Count: 1
Battery/Capacitor Status: OK
Michael A. McKenney
Respected Contributor

Re: Smart Array 6i - very poor read performance

Did you test it with a benchmark program like Sandra Pro?
GregLS
Occasional Visitor

Re: Smart Array 6i - very poor read performance

> Did you test it with a benchmark program like Sandra Pro?

No, just tested using standard tools like hdparm and dd.

Some additional information here:

1 - Moving from the CentOS kernel to a more recent kernel.org kernel increased the speed from 40MB/sec to 50MB/sec.

2 - Using --direct on hdparm increased the transfer rate from 50MB/sec to over 80MB/sec, sometimes as high as 120MB/sec. I have never seen that kind of difference with and without O_DIRECT on a system before, but the O_DIRECT speeds are on target with what I was expecting.
Michael A. McKenney
Respected Contributor

Re: Smart Array 6i - very poor read performance

Upgrading the kernel and drivers might help performance. Did you verify the firmware revision to make sure you have on conflicts.