ProLiant Servers (ML,DL,SL)
1753415 Members
7105 Online
108793 Solutions
New Discussion юеВ

Re: Migrate band size RAID 5

 
Robertomcat
Regular Advisor

Migrate band size RAID 5

Hello, good afternoon.

I wanted to ask you what could be the best optimization and efficiency for an 8├Ч6 TB RAID5 array.

Currently I have this band size, and now I'm going to dedicate this RAID to download through P2P, so it will have an intense work of reading, but less than 80/20 writing.

I've been looking, and there are people who say to lower the band size to KiB, but I would like to know your opinion in order to optimize the RAID as much as possible.

Thank you!

5 REPLIES 5
BPSingh
HPE Pro

Re: Migrate band size RAID 5

When distributing data across multiple physical drives (striping), the strip size is the amount of data that is written to each physical drive. The full stripe size refers to the combined size of all the strips across all physical drives, excluding parity-only drives.

The maximum strip size changes dynamically and is reduced for arrays with a large number of data drives or with smaller controller cache sizes. The controller must be able to read an entire strip of data at a time into cache memory during transformation. Available memory is the limiting factor and also check on the read/write settings on the cache module.

Another thing that we should keep in mind is the about the job types (sequential or random). For random jobs then it is advisable to keep the stripe size large but keep it small for sequential jobs.

 


I am an HPE employee.
[Any personal opinions expressed are mine, and not official statements on behalf of Hewlett Packard Enterprise]
Accept or Kudo
Robertomcat
Regular Advisor

Re: Migrate band size RAID 5


@BPSingh wrote:

When distributing data across multiple physical drives (striping), the strip size is the amount of data that is written to each physical drive. The full stripe size refers to the combined size of all the strips across all physical drives, excluding parity-only drives.

The maximum strip size changes dynamically and is reduced for arrays with a large number of data drives or with smaller controller cache sizes. The controller must be able to read an entire strip of data at a time into cache memory during transformation. Available memory is the limiting factor and also check on the read/write settings on the cache module.

Another thing that we should keep in mind is the about the job types (sequential or random). For random jobs then it is advisable to keep the stripe size large but keep it small for sequential jobs.

 


Hello, thank you for your reply!

As for the configuration of the hardware controller (P440/4GB), currently all the cache is assigned to write, nothing to read. The server has a RAM of 32GB.

On sequential or random readings, most readings will be random, so as you say, I think the band size is correct.

If you have any more ideas on how many configurations, please let me know. Thank you!

Robertomcat
Regular Advisor

Re: Migrate band size RAID 5

@BPSinghAnother question that comes to mind. Do you know if the Windows native defragmenter would do a good job defragmenting RAID5 8├Ч6TB hardware?

BPSingh
HPE Pro

Re: Migrate band size RAID 5

 

The controller cache size is 4 GB. The read/write settings on the cache can be adjusted from Smart Storage administrator, you may set it to 25% read and 75% write or 50/50 depending on your requirement. Please be informed that the system memory (32GB) won't be used for this.


Disk Defragmenter is known to increase data access speeds but however not sure about its affect on RAID performance. You may have to check with the OS vendor for this.


I am an HPE employee.
[Any personal opinions expressed are mine, and not official statements on behalf of Hewlett Packard Enterprise]
Accept or Kudo
Robertomcat
Regular Advisor

Re: Migrate band size RAID 5

Ok, thanks!!!!!!