Subscribe

Performance Manager Volume Latency

I'm new to c-mode, unified manager, & perf. manager.

 

I have a FAS2552.

I'm running 8.3.2

 

I have one volume (approx 200GB) used for NFS on the cluster with not many IOPS.

 

I'm getting latency alerts in performance manager for the following policy:

-------------------------

Object Type: Volume

Object Counter Condition* Avg Latency ms/op Warning: 20 Critical: 30

 

Thresholds must be crossed for at least* 5 Minutes

----------------------------

 

The volume is reporting critical errors because the latency ms/op for this volume averages 274 with a high of 743.

 

I break down the details and discover that the source of the latency is being caused by the 'Network Processing' Cluster component.  All other components are contributing 0 to the latency.

 

My read/write latency is < 0ms 

 

IOPS avg around 100.

 

 

Thoughts?

 

(CPU isn't doing anything either.)