HPE Community read-only access December 15, 2018
This is a maintenance upgrade. You will be able to read articles and posts, but not post or reply.
Hours:
Dec 15, 4:00 am to 10:00 am UTC
Dec 14, 10:00 pm CST to Dec 15, 4:00 am CST
Dec 14, 8:00 pm PST to Dec 15, 2:00 am PST
MSA Storage
cancel
Showing results for 
Search instead for 
Did you mean: 

4KB boundaries and proper block size for SSDs

 
SOLVED
Go to solution
Highlighted
JustinN1
Occasional Contributor

4KB boundaries and proper block size for SSDs

When reviewing the MSA 2050 SMU reference guide I noticed this tidbit in relation to SSDs and write amplification.  It states that

Data that is written in sequential LBAs that are aligned on 4KB boundaries results in the best write amplification factor. The worst write amplification factor typically occurs for randomly written LBAs of transfer sizes that are less than 4KB and that originate on LBA's that are not on 4KB boundaries. Try to align your data on 4KB boundaries

When it states "4KB" boundaries, is it referring to a block size of 4KB?  Does that mean that the block size, or "allocation unit size" as Windows calls it should be 4KB?  Or would a multiple of 4KB, such as 8KB, 64KB, etc, work?  I'm looking to build a few archive LUNs with archive affinity for our larger file servers that just have data sitting there, and with 4KB block size I can only get the LUNs up to 16TB in Windows.  I know that with archive affinity it shouldn't write to the performance tier at all, but I may at some point want a larger-than-16TB volume for performance data as well.

3 REPLIES
Solution

Re: 4KB boundaries and proper block size for SSDs

I would suggest you to refer the below link and check "Sector format" and "Chunk Size",

https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-c05075467

I would suggest to keep in mind when creating Virtual Disk Group, number of drives and Chunk size selection. Please find some important notes from SMU guide,

"NOTE: For a virtual group, the system will use one of the following chunk sizes, which cannot be changed:
– RAID 1: Not applicable
– RAID 5 and RAID 6:
• With 2, 4, or 8 non-parity disks: 512k. 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.
• Other configurations: 64k
– RAID 10: 512k"

 

So yes the size of data that you are transferring from Host is important to decide on chunk size on MSA. This is why we say data alignment should be 4KB boundaries.

 

Hope this helps!
Regards
Subhajit

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

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

Re: 4KB boundaries and proper block size for SSDs

Can you please help us to understand if the issue got resolved or not?

If issue got resolved then how?

This will help for everyone who are all following your forum.

 

Hope this helps!
Regards
Subhajit

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

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

JustinN1
Occasional Contributor

Re: 4KB boundaries and proper block size for SSDs

Sorry, I'm having a bit of difficulty with this website.  It looks like I'm finally able to login and type in the text box (it kept disappering).

This information seems to be the most pertinent that I have found.  I've been settling on setting the block size for Windows-based LUNs to the default 4k and the block size for ESXi to the only selectable option in the drop down (1MB).