average disk queue length at 100 percent

[barnegattech]

I am running server 2003. The server is dogging. I have a average disk queue length pinned at 100 percent. Also the page/sec is high. It usually pins at 100 when the disk queue length does. There seems to be a bottle neck in the area of I/O. Can something be done to eliminate it.

 

[technome]

Please supply more info..
Server model/speed

Raid equipped, type(raid 1, raid 5)adapter model (Perc 4,5,etc)how many disks, disk type(SAS, SATA, SCSI), if raid raid policy (write through or write back).

How many users accessing, light or heavy load.

Could network infrastructure be causing delays? quality of the wiring, 100 or gig switches/clients.

What programs are being used, server resident and client to server access.


........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[technome]

How much ram?


........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[itsp1965]

To add to the above what apps are you running (is there a virusscan running)? I have seen SQL Servers that were perpetually paging on a system with 32GB of RAM. Try checking task manager (not the best tool) and enable the additional counters dealing with paging, pagefile it should be able to identify your culprits.

 

[barnegattech]

server is an custom build with all intel
BX80563E5335P Quad-Core Intel® Xeon® processor Xeon E5335 2GHz 1333MHz 8MB (Two processors)
SRCS16 Intel SATA RAID Controller
4GB ram
4 sata hard drives
not many users but a production machine with high I/O; scanning stations transferring scan documents to the server. I would consider it a heavy load
giga bit network. Network traffic is to be ok
Virus scan is not running and disabled sql server from service with no change.
maybe the page/sec being high is fine or even good. but the average disk queue length I don't believe should be high.

 

[barnegattech]

2 drives 500GB each on partitioned system/data on raid 1
2 drives Seagate 1.5TB hard drive 7200RPM 32GB cache 3GB/s each for high production data on raid 1.

 

[technome]

What is the raid cache policy, "write back" give a decent performance boost even in a raid 1.

http://www.intel.com/support/motherboards/server/sb/cs-011129.htm

Could be many reasons for the queue length, not necessarily the disks, as delays caused by network/programs issues.
If you have managed switches are the ports error free or have a low number of errors?
I had major slow downs with SMB signing enabled, but generally along with very pronounced network stability issues.
All the above assumes you are malware/virus free.




........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm