EvaPerf counters not reliable or bad interpretation of their meaning ?

[Max Pierre]

Hi,

Eva P6300, EvaPerf 10.2.0.091012A

We monitor the arrays with the evaperf command :

evaperf mof -sz p6300a p6300b -csv -cont 60 -dur 3600 -od c:\evaperf

When we compare hps and pdg  we found very different values.

For example, if we "add" all FP ports Write MB/s for one controller from hps , the result is about 3X higher than adding all average Write MB/s on the same controller for all disk groups.

Sum of mean MB/s values for FPn on ctrl A (hps) =~ 3 x Sum of mean Average Write MB/s for all disk groups on ctrl A (pdg).

Why a such difference ?

Any explanation is welcome

P.S. From hps, we take into account FP1 and FP2 only as FP3 and FP4 are dedicated to data replication