I have a VMAX 250FX running PowerMax OS with Unisphere 9 and SE 9. I have noticed the compression ratio is different among the GUI, and CLI so I get confused which ratio should be used.
1. The GUI show data reduction of 2.8:1
2. symcfg list -srp -efficiency show 2.6:1
3. symcfg list -pool report compression ratio of 2.2:1
If I want to calculate capacity saving before and after apply compression, which output I should rely on.
Not sure why that would be, but you can verify your numbers based on the pools (provided you don't have any DSE tracks allocated).
The compression ratio for a pool is the ratio of the sum of logical used pool capacity and sum of physical pool capacity (excluding DSE tracks).
Logical pool capacity is sum of tracks used * track size of 128.
Physical pool capacity is sum of used tracks * pool track size.
You can determine the pool track size by a symdev show of a TDAT in the particular pool. Take the sectors per track number / 2 and that is the pool track size.
Note that if a pool is empty, it will show the maximum compression ratio available for that pool. for example an empty 32K track size pool will show up as 4.0:1
If you do the math on your pools, does it match the pool report?
For each pool, calculate logical tracks used by taking tracks used * 128.
For each pool, calculate Physical tracks used by taking tracks used * Pool_track_size
Sum The physical numbers and sum, the logical numbers.
Divide the total Logical by total physical and that is the overall compression ratio.
The above calculation on your data comes up with a ratio of 2.1:1 Exactly the number in the totals line. So I would say that the 2.1:1 is accurate based on the numbers in the pools at the time the command was run.
I noticed that the "symcfg -sid list -srp -efficiency" will return data reduction ratio of 2.6:1. I am not sure how this 2.6:1 is being calculated. The array is VMAX 250FX (VMAX AFA) with PowerMax OS 5978.144.144 so in this case do we have deduplication turn on. I thought deduplication is only available for PowerMax array with PowerMax OS so the data reduction ratio should return the same ratio in VMAX 250FX because we only have compression. For capacity planning purpose, which compression ratio I should be used in this case.
PedalHarder
4 Apprentice
•
465 Posts
5785
0
Posted August 6th, 2018 19:00
Not sure why that would be, but you can verify your numbers based on the pools (provided you don't have any DSE tracks allocated).
The compression ratio for a pool is the ratio of the sum of logical used pool capacity and sum of physical pool capacity (excluding DSE tracks).
Logical pool capacity is sum of tracks used * track size of 128.
Physical pool capacity is sum of used tracks * pool track size.
You can determine the pool track size by a symdev show of a TDAT in the particular pool. Take the sectors per track number / 2 and that is the pool track size.
Note that if a pool is empty, it will show the maximum compression ratio available for that pool. for example an empty 32K track size pool will show up as 4.0:1
If you do the math on your pools, does it match the pool report?