Originally posted by: Aruna
We are seeing high I/O latency spikes in our ESXi 5.0 farm from the vCenter for LUNs provisioned from an AMS2300. Attached here one sample screen of I/O latency spikes. The setup is AMS2300 <-> MDS 9124 <-> Cisco UCS 5108.
[attachment=28:Physical device command latency - Screen_1.JPG]
AMS was found to be healthy and no resource bottleneck by looking at PFM data. Cisco level also all buffers/logs are OK as per respective vendors.
Any one seen this type of behavior before ? appreciate any advice finding a resolution.
I/O latency spikes in the vCenter varies from 100ms to 1000ms range and frequent in nature. There is no association to a pattern or workload also, identified so far.
Thanks in advance.