Performance analysis example

Analyze response time

1. Check performance of the affected storage resources:

   1. See performance metrics for parity groups, MPB, I/O paths, response time, and cache write pending.
   2. If the storage system is modular, see performance metrics for UVM/VM response time, HDD operation rates, CPU core busy, cache write pending.

2. Consider the following components:

   1. HBA queue depth
   2. HBA settings
   3. Virtual queue depth settings on the ESX host

3. Check the following switch components:

   1. Switch configuration
   2. Buffer-to-buffer credit hits
   3. ISL links
   4. Hardware failures (for example, an SFP module might not be properly connected)

Analyze IOPS

1. Check for workload issues.
   • Is the storage tier type incorrect?
   • Is the profile for this application correct?
   • Is there enough disk space to handle the workload?
2. Check the performance for Parity Groups, MPBs, I/O paths.

Check utilization

If throughput is high:

1. Check all associated data paths for high utilization.
2. Check the CHA>ESX host paths.