VMware: “adaptive queue depth” setting

What is the recommended LUN queue depth throttling in VMware ESX/ESXi?

According to EMC this “adaptive queue depth” setting is not yet supported (June 2013). The full article can be found on this link. It is known as Primus solution id emc279718.

My advice would be not to use this setting until EMC certifies this setting for use with their arrays.

The exact Article is as follows:

What is the recommended LUN queue depth throttling in VMware ESX/ESXi?
 
  • Need to know QFullThreshold,QFullSampleSize & Queue Fulls values for VMware.
  • Host Connectivity Guide does not include any standard values.
  • Performance issues.
 
OS: VMware ESX / ESXi
Product: CLARiiON / Symmetrix and I assume VNX as well
 
According to Engineering, enabling the adaptive queue depth algorithm is NOT documented in the Host Connectivity Guide or the EMC Support Matrix.
The recommendation is to use the default setting which is disable adaptive queue depth.
Additional checks with E-Lab VMware quality engineers, E-Lab testing is done with the default setting.
 
See emc274169 for additional information and or changes.  Also see these two documents on support pages support.emc.com or Powerlink.Using EMC VNX Storage with VMware VsphereUsing EMC Clariion Storage with Vmware Vsphere & VMware Infracture Version 4.0

Would you like to comment on this post?