Jaa

Two Blog Follow ups

  • Artikkeli

I have two blog subject follow-ups tonight:

Follow-up #1... Regarding my "Page Cleaning with sp_clean_db_file_free_space and sp_clean_db_free_space" post - my friend and colleague Jimmy May let me know that the always amazing Kalen Daleney just posted on this subject today which you can read here: Clean Up Your SQL Server Databases.  One good point that I didn't mention in my post but that Kalen did - you can use these procedures to compensate for the potential Instant File Initialization security considerations. 

Follow-up #2... Regarding my post "SQL Server and HBA Queue Depth Mashup" - I had a chance to test out HBA Queue Depth settings recently and had unexpected results.  First of all, I used SQLIO against a QLogic HBA and EMC SAN.  In my testing scenario I was primarily interested in write-performance with varying queue depths.  I tested random 8K writes, sequential 64K writes, and sequential 256K writes using depths of 16, 32, 64, and 96.  The result?  For random 8K writes, 32 was the best setting by far (for example - issuing 83 MB/sec versus 47 MB/sec for a 64 queue depth).  I found that a 64 queue depth was only marginally better for sequential 64k writes and that for 256k sequential writes a 32 queue depth was significantly better than a 64 queue depth.  So the default queue depth of "32" for the Qlogic HBA worked best for random 8k and sequential 256k writes.  Not what I expected (was hoping to eek out a little more throughput with a higher queue depth) - but that is why we test, no?  There are several factors involved, so always test for yourself given your own hardware components, topology, and environment.

Happy Thursday night!

Joe

Comments