Hi Guys

I would like to understand more about the best practices for the Lun design keeping in mind the dynamic provisioing and virtual luns carved from the pools.

In earlier times we used to make sure we have many smaller luns for e.g 4x25GB assigned for better performance with Raid Groups

, but with dynamic provisioing and pools in picture with Vmax , we believe that providing 100GB one Lun would be much better than 4x25GB , since they use the same disk at the backend. But still 4x25GB will do.....

1. But does it make more sense to have 100GB single lun compared to 4x25GB lun to a FS which is concantenated ?

Normally we talk about performance in IOPS and calculate IOPS per GB , so If for example , the application team comes to me and requires 5000IOPS , then I will calculate the storage i need to allocate as per the set value for example (imaginary)  100GB on standard will give  1000IOPS , so  would require 500GB to achive the IOPS requirement ?

But if I give the storage as multiples of 50GB , would it cause a performance problem since the FS is concantenated and only first chunk of 50GB will used for RW/R operations initially ?

2. Another question is on , how we should manage the volume group based on Data/redo/archive FS on the dynamic provisioing on the Pool ?

I Presume if I have two Pools I can provide separate disk from each pool for data/redo But this pool will be used for other servers as well, whose data disks would sit on the same backend spindles on which redo/archive IO are happening , so does it really make any sense ?

What are the best practices ?