HPE StoreVirtual Storage / LeftHand
cancel
Showing results for 
Search instead for 
Did you mean: 

Queue Depth Total Explanation

Nick Salty
Occasional Advisor

Queue Depth Total Explanation

Hello,

I have a P4500 G1 4-node  multi-site cluster.

I am trying to find where a performance bottleneck is for my backup process.

I use the P4500 for iSCSI storage from both VMware and Microsoft Windows

 

None of the network connections look to be more than about 20% utilised, so now I'm looking at the storage.

 

Looking at the Queue Depth Total (QDT) count, I don't think I understand what it is telling me:

 

If I select the Cluster, then the QDT is 36,830 (current, max and average are about the same)

but if I select to see each node, then QDT has a max of 6 across all the nodes (actual maximums are: 4, 5, 1, 6)

 

So why would the cluster value be crazy-high, when each node is perfectly reasonable?

 

Thanks for any tips,

N

2 REPLIES
Emilo
Trusted Contributor

Re: Queue Depth Total Explanation

This is an known issue with and older version of the CMC, it is only cosmetic as when you did the individual nodes it showed normail values. Update the CMC and see if the cluster values  = the sum of the individual nodes.

The bug was in one of the first releases of the CMC.

Normal queue depth totals are  2 x # of disks in in the cluser for SAS and 1 x # disk in the cluster for sata.

 

bryan_hadzik
Frequent Advisor

Re: Queue Depth Total Explanation

Emilo is right, that is a bug in the CMC. Get the latest version, it is backwards compatible. The number should be pretty low. Tell us more about your setup. Are you using the DSM for MPIO? What is your IOPS cluster wide?
Do you have sata/mdl sas/sas disks?