Storage Boards Cleanup
To make it easier to find information about HPE Storage products and solutions, we are doing spring cleaning. This includes consolidation of some older boards, and a simpler structure that more accurately reflects how people use HPE Storage.
Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

SCSI Queue Depth VA7400

SCSI Queue Depth VA7400

I have read a lot about tuning the SCSI queue depth to be higher than 8 when you have a single large LUN holding an Oracle database. Given that default param is 8, can someone tell me what happens to I/O when the Queue Length of 8 is exceeded.
Regards
1 REPLY
Eugeny Brychkov
Honored Contributor

Re: SCSI Queue Depth VA7400

I guess application execution is suspended until there will be a slot for IO in the queue.
Please remember that in addition to queue at the host side (in the SCSI driver) there's queue and tunable queue depth at the disk array side. And 'optimal' queue length (depth) setting depends on how many disks you have in disk array.
If looking to disk array side: too long queue causes disk array to take too many requests and some times requests are not processed fast enough and got aborted by the host. If queue is too short then disk array takes too few IOs and responds 'busy' to another host outstanding IOs. Generally too long queue is better than too short.
Almost the same with local SCSI driver queue.
As soon as in hpux each LUN can have its own queue depth but at disk array side there's only one queue, then the ideal is to have sum(q_depth(LUN#))=disk_array_q_depth. Of course this formula depends on how many host ports disk array has, but idea is clear
Eugeny