MSA 2052 tiering

[homermg]

Hi all,

IM just trying to understand my new MSA 2052.

So i have just created a raid 1 with two SSD'S within Pool A and Raid 5 Diskgroup - Volume also within Pool A ( HD 2,5 10000rpm)

Us i understand the best practise evyrthing within the same Pool will use automaticly tier performance.

1. Waht i not really understand do i have to selct Tier Affinity: performance on the volumes?

and

2. Is there a chance to see if tiering is running? somwhere on performance tab maybe?

and

3. how to understand the performacen tab, i can't find any info about that.

Hope you can help me.

Many thansk

to all

[VikasMandal]

Dear @homermg

Thank you for your questions.

The answer to all your questions can be found in the best practice guide.

Volume Tier Affinity is a settable attribute that allows a storage administrator to define quality of service (QoS) preferences for virtual volumes in a tiered environment. There are three Volume Tier Affinity options—Archive, Performance, and No Affinity. A setting of Archive will prefer the lowest tier of service, Performance will prefer the higher tiers of service, and No Affinity will use the Standard Tiering strategy. Tier Affinity is not the same as Tier Pinning and does not restrict data to a given tier and capacity. Data on a volume with Archive affinity can still to be promoted to a Performance Tier if that data becomes in demand to the host application.
Refer Best Practice: https://h20195.www2.hpe.com/v2/Getdocument.aspx?docname=a00015961enw

The tier I/O panel in the footer shows a color-coded bar for each virtual pool (A, B, or both) that has active I/O. The bars are sized to represent the relative IOPS for each pool. Each bar contains a segment for each tier that has active I/O. Hover the cursor anywhere in this panel to display the Tier I/O Information panel, which shows the following details for each tier in each virtual pool:
• Current IOPS for the pool, calculated over the interval since these statistics were last requested or reset.
• Current data throughput (MB/s) for the pool, calculated over the interval since these statistics were last requested or reset.
The panel also contains combined total percentages of IOPS and current data throughput (MB/s) for both pools. 

HPE MSA 1040/2040/P2000 G3 Systems - Performance Monitoring via SMU (Understanding performance tab):
https://support.hpe.com/hpsc/doc/public/display?docId=mmr_kc-0119548

Please refer best practice guide for more information:
https://h20195.www2.hpe.com/v2/Getdocument.aspx?docname=a00015961enw

Hope this helps.

Regards,
Vikas

[homermg]

Hi Vikas,

im still not sure what it means?

As i understand i need just to create a raid 1 with my both ssd's disk types: esas performance on e.g. Pool A.

Than evyrthing i craets on pool A will uses the ssd's for performance tiering without to selcet any affinity.

Is that correct?

So affinity need is only to control  that if e.g. the Pool A with the SSD's has several volumes but i don't use the SSD's for all the volumes, is that right?

Thanks

[SUBHAJIT KHANBARMAN_1 · Accepted solution]

Yes your understanding is correct

When we have tier model then any volume part of that Pool will be using tier architecture until and unless we have any specific rule set as Volume Tier Affinity

now if you have any specific application requirement for example let's say you have any backup application or something which is not so important and you want these data always should go to your Archive tier - in that case you should assign your Volume Tier Affinity for that volume as "Archive" so that these 4MB pages of data will never touch SSD which may be require for handling some important production data.

So Volume Tier Affinity is volume specific and this is not applicable to any volume part of that Pool. Other volumes will be using 3 Tier Architecture of that Pool as it is.

Hope this helps!
Regards
Subhajit

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

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

[homermg]

cool many thanks! got it :-)

[arora_newtech]

what is best throughput is expected from MSA2052 with 30 x 12tb 7.2k rpm disks

[SUBHAJIT KHANBARMAN_1]

Request you to kindly open new discussion forum with your new query as we need to address each issue seperately so that everyone can track the same based on topic.

I would suggest to check page no 9 onwards topic "Performance" from quickspec that will give you idea of what all inputs required before we can calculate approx throughput which you are asking,

https://h20195.www2.hpe.com/v2/getdocument.aspx?docname=a00008277enw

Hope this helps!
Regards
Subhajit

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

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