Accepted Solution

Where do all the aggregate IOPS come from?



In my OCUM, I open the graph showing IOPS overtime for an aggregate, then click and include each and every volume in that aggregate....

I expect the aggrerate IOPS to be very close to the sum of the IOPS of the contained volumes, but it does not seem to be the case!

The aggrate IOPS are in the 10k, while the sum of the volumes is more like 1,2 or maybe 3k at best.


What am I missing? Thank you so much for any pointers, as this question is driving me crazy!



Re: Where do all the aggregate IOPS come from?



It may be due to bug:

If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.

Re: Where do all the aggregate IOPS come from?

The aggregate IOPS represent the IOS served by the aggregate which are not equal to the IOs requested by the user due to multiple processing that happens on the way, such as 


1 - large user IOs are divided into smaller IOs to be served by the aggregate

2 - system reads more data from the aggregate in order to improve cache hit rate for the user traffic

3 - data is read/written in aggregate to comply with protection features such as  clone/backup/snapmirror etc.


In order to determine which applies in this case, a deeper review of the system needs to happen.