Unsolved
This post is more than 5 years old
36 Posts
0
1398
What is the best sized lun from a VNX pool
Hi,
What is the best sized lun from a VNX pool ?
How the luns are getting distributed on the VNX storage pool . We have Thick Luns.
Please share your views and inputs..
kelleg
4.5K Posts
0
October 23rd, 2015 12:00
There is no "ideal" size for a LUN. The size of a LUN will depend on your needs for capacity and performance and what you want to the use the LUN for.
I recommend that you read the White Papers for Virtual Provisioning for VNX2 and VNX2 Unified Best Practice Performance. There are available on the support.emc.com site under Support By Product. Select your array model (VNX1 Series or VNX2 Series) and then look for the White Papers.
Introduction to the EMC VNX2 Series - A Detailed Review
VNX2 Unified Best Practices for Performance - Applied Best Practices Guide
Virtual Provisioning for the VNX2 Series - Applied Technology
Introduction to EMC VNX2 Storage Efficiency Technologies
glen
Anil2sharma
36 Posts
0
October 25th, 2015 09:00
Hi Glen,
Thank for your inputs.
Yes, I also have the same view point that there is no Ideal LUN size when we takl about the storage POOL lun.
But, can you confirm the below points which is as per my understanding..
1) When we create a lun in a pool lets say of size 100 GB. It distributed on the drives of luns in the slices of 246mb .
So that it can get the performance of all the drives.
2) And if my above understanding is right, then its not matter its a small lun or big size lun in a pool.
Is this correct. ?
Rainer_EMC
8.6K Posts
0
October 25th, 2015 14:00
For use with data mover (VNX file NAS) there are recommendations in the best practice paper
Rainer_EMC
8.6K Posts
0
October 26th, 2015 06:00
256 MB is the slice size but blocks are allocated at 8k
yes we try to distribute LUNs across a pool but its not a simple striping - depends on space available, FAST, thick vs. thin
recommendations about the LUN size are most often because of client performance considerations - like having multiple I/O queues with more outstanding I/O's can have an advantage.