There is a set of algorithm that is run for every type of Request, Disk type, Storage Pool and Raid Type.
The algorithm will make a performance pool give more performance and compromising storage space and a Storage Based pool will be based on high storage but compromising on Performance.
In VNXe it will lay down all the FS on all the Disks from which the storage is coming from.
In your case considering the disk size to be 100GB
2 Raids:
RAID-5 6+1
RAID-5 6+1
So the total Space is 600 + 600 Approx.
So If you create 6 Equal LUNs from the Raid
So you have 12 LUNS coming from 2 Raids and each LUN is a Strip of all 7 Disks in the Backend(I mean Raid)
When you request a FS got 400 GB it is gone Take 2 LUNs from the First raid and the other from the Second raid, so that all the 14 Spindle act for an action on that FS and also Both SP are engaged for a better performance.
I have attached the screen shot of a Small algorithm for clar_r5_performance storage pool. The volumes in the storage pools are balanced between SP A and SP B.
Kumar_A
2 Intern
•
727 Posts
0
June 3rd, 2011 09:00
The LUN would be laid out on all the drives in the pool.
lenystr
43 Posts
0
June 7th, 2011 00:00
How is it laid out (i.e. how many blocks per disk in the stripe)? -Thank you
rajnvr1
57 Posts
0
September 29th, 2011 04:00
There is a set of algorithm that is run for every type of Request, Disk type, Storage Pool and Raid Type.
The algorithm will make a performance pool give more performance and compromising storage space and a Storage Based pool will be based on high storage but compromising on Performance.
In VNXe it will lay down all the FS on all the Disks from which the storage is coming from.
In your case considering the disk size to be 100GB
2 Raids:
RAID-5 6+1
RAID-5 6+1
So the total Space is 600 + 600 Approx.
So If you create 6 Equal LUNs from the Raid
So you have 12 LUNS coming from 2 Raids and each LUN is a Strip of all 7 Disks in the Backend(I mean Raid)
When you request a FS got 400 GB it is gone Take 2 LUNs from the First raid and the other from the Second raid, so that all the 14 Spindle act for an action on that FS and also Both SP are engaged for a better performance.
I have attached the screen shot of a Small algorithm for clar_r5_performance storage pool. The volumes in the storage pools are balanced between SP A and SP B.