We are trying to run some simple capacity reporting as follows:
-capacity in each tier
-raw capacity in each tier
-total capacity of each array
-total raw capacity of each array
Our environment has NetApp, HDS, and AMS HDS arrays being front ended by HDS VSP's and V-series. We are wondering what the recommended data model would be for our reporting needs.
In the past we have used the Capacity>Storage Capacity data model (using Current Capacity Fact), but the 6.4 Reporting guide seems to suggest using the Capacity>Storage and Storage Pool Capacity>Physical Capacity data model for these types of reports (using a date or the like to get only current data).
The output between the two reports is different enough to question which one is the correct data model to use in our environment.
Hoping someone out there has some words of wisdom for our situation. Thanks for any help/advice.
Looking at the question that way makes it hard to provide a meaningful answer.
If I were writing the report you describe from scratch, I'd use Public Folders > Capacity > Storage and Storage Pool Capacity. And you would want to use Physical Capacity inside of that mart.
However, you already have a capacity report. Is its output correct? If so, why would you re-write it on another data mart? Is it incorrect? Then you might want to think about how much effort went into producing it, and how hard it would be to start from scratch. Do you wonder whether it's correct? The best way I've found to figure that out is to break the report down into a smaller test case (one array, for instance) and compare it to the data from your favorite array management tool. This works particularly well if you've structured your reports to roll up from much detail to a summary level of detail.
We had some situations where the reports from the different data models were not matching. We have a diverse environment with a lot of legacy hardware and tools (some bad, and some not so good). So as you can imagine, it can be a bit tricky and tedious to calculate at the Tier level (just one of the reasons OCI has been such a great help in this environment). Before we went too far down that road, we wanted validate whether there was a recommended data model for our situation. Sounds like it just depends…
We are doing some testing/calculations now, and I will report back what are results are...