This is from a customer situation
in the week from September 17, 2012 severe performance issues have been detected on several hosts in the VMware Farm.
A first analysis showed that one of the responsible discs has been ESX_D37.
The problem was especially high on Monday from 10 AM to 1 PM.
We want to analyze the situation on Monday to find out with the problem was coming from
This Analysis will cover the following
look into the performance of a volume - very simple - only volume IOPS and volume response times but good enough for this purpose
analyze managed disk group cache behavior - the mdg cache is total overfilled - we see extrem high cache page destage operations
volume analysis of all volumes in the managed disk group to find which volumes are responsible to fill up the cache
additionally with a look into the cache usage of these volumes
Link to the Video