Chunk size and type of disk group on MSA2040

[abelono]

Hi all!
What is the difference between a linear pool and a virtual pool on msa2040? Is there a big difference in the types (linear or virtual) of disk groups in terms of performance?
The vendor has recommendations on chunk sizes for applications, for example, Microsoft SQL (64KB-512KB)? The wrong chunk size can also lead to performance degradation?
Thanks.

[SUBHAJIT KHANBARMAN_1]

@abelono 

In order to understand difference between Linear and Virtyal array for MSA 2040 I would suggest to go through SMU guide and best practices guide to get good amount of information.

https://support.hpe.com/hpesc/public/docDisplay?docId=c04220794

https://www.hpe.com/psnow/doc/4aa4-6892enw?jumpid=in_lit-psnow-red

Few things to mention that, you will not get option to select Chunk size in case of Virtual Disk Group (VDG) creation because sytem automatically decide Chunk size based on RAID selection and number of drives. For example, a RAID-5 group with 3, 5, or 9 total disks or a RAID-6 group with 4, 6, or 10 total disks, Chunk size 512k. Also for RAID 10 its 512k only.

As per best practice guide, for SQL Application RAID 5 should be for data and RAID 10 for logs. Read-ahead Cache size need to be Adaptive and Cache write optimization should be Standard.

In order to get good performance you need to keep in mind that your MSA Chunk size should match with IO block size of the application.

Virtual storage offers a great deal of functionality enhancements, there is an overhead associated with it. So there is no comparison between Linear and Virtual array.

Automated Tiering uses virtualized storage and is accomplished by paging. MSA virtual volumes allocate data into 4 MB chunks (pages) from within the virtual storage pool. The pages are ranked based on a sophisticated algorithm. The page rank is used to efficiently select appropriate pages to move between tiers. The result is that pages can be migrated between tiers automatically such that I/Os are optimized in real-time.

Another major thing to conside is Sequential or Random IO for database.

Hope this helps!
Regards
Subhajit

I am an HPE employee

If you feel this was helpful please click the KUDOS! thumb below!

**************************************************************************

[Sebasbin]

HI,

Kindly go through below SMU Reference Guide and best practices to get the information about linear and virtual pool

https://support.hpe.com/hpesc/public/docDisplay?docId=c04220794

https://www.hpe.com/psnow/doc/4aa4-6892enw?jumpid=in_lit-psnow-red