Dedupe vVOL

 
Sanja1
Frequent Advisor

Dedupe vVOL

We are having Primera 630 with all SDD drives with one vVOL. I was given instructions because we have vVOL and no VVs, to enable dedupe inside vsphere. Under storage policy I enabled rule for CPG- data reduction.

I can see that dedute is enabled on VMs (on ones i applied the strage policy) but I am still confused who is doing dedupe here...primera or vmware? and how does it works? 

7 REPLIES 7
sbhat09
HPE Pro

Re: Dedupe vVOL

Hello Sanja,

Good question.!
It is actually Primera doing data reduction and VMware supports it. Data reduction is the combination of dedupe and compression that we see in 3PAR. In simple words, the inbuilt algorithm of Primera's ASIC identifies redundant patterns of data writes in multiple locations of the volume. Remove those redundant entries to save space without losing any data for the user.

Data reduction is ideal for volumes that have a high level of redundant or compressible nonredundant data. Data sets that are good candidates include:
• Virtual machine images
• Virtual desktop infrastructure
• Databases
• Home directory and file shares

Data sets with a low level of redundancy are not good candidates. For example:
• Compressed data
• Encrypted data

For more details, please refer the doc (pg no 15) - https://psnow.ext.hpe.com/doc/a50000189enw 

Regards,
Srinivas Bhat

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


I am an HPE Employee

Accept or Kudo

Sanja1
Frequent Advisor

Re: Dedupe vVOL

Thank you for your answer. One more doubt...

Primera shows no Data reduction savings

Screenshot_57.jpg

Still we I go on SCCM and my Container I can see Dedupe savings..

Screenshot_1.jpg

 

sbhat09
HPE Pro

Re: Dedupe vVOL

Hi Sanja,

The first screenshot is from Primera UI for a specific Primera system. Total host written data is only 100 GiB. There could be several volumes in the Primera system with very little data written on them. Primera's data reduction (compression + deduplication) works within the volume. I think within those volumes, there may not be qualified data for data reduction, or data reduction is not enabled in many of them.

The second screenshot is from SSMC. By default, SSMC shows the details of all the 3PAR and Primera systems managed from it. The capacity efficiency data showing here looks like is for all the 3PAR/Primera systems.

Regards,
Srinivas Bhat

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


I am an HPE Employee

Accept or Kudo

Sanja1
Frequent Advisor

Re: Dedupe vVOL

We only have one container where al vVOLs are stored

sbhat09
HPE Pro

Re: Dedupe vVOL

Hi Sanja,

But I see only 0.1 TiB host written data in the first screenshot. And the utilization looks much higher in the second screenshot. Can you please confirm that?

Regards,
Srinivas Bhat

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


I am an HPE Employee

Accept or Kudo

Sanja1
Frequent Advisor

Re: Dedupe vVOL

I'm little bi confused. only VV (600Gb) is one containing vcenter. could it be that?

sbhat09
HPE Pro

Re: Dedupe vVOL

In SSMC, go to Main menu -> Virtual volumes. 

Check how many volumes are there ( on left-hand top corner of the screen, like 1st screenshot) and if it is 'All systems' selected (on right-hand top corner of the screen, like 2nd screenshot).

Snj1.JPG Snj2.JPG

In Primera, anyway it will be only one system. If the number of systems and volumes are the same on both Primera UI and 3PAR, then the screenshots you sent today are comparable.

I think the capacity details from SSMC is for multiple 3PAR/Primera systems and more volumes. The capacity details you see in Primera is for single Primera system and fewer volumes.

I am telling this based on difference in used capacity in Primera UI and SSMC. Sorry if I am wrong.

Please let me know if that resolves your doubt. Else, I may ask you to send some specific details to understand more and try to assist further.

Regards,
Srinivas Bhat

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


I am an HPE Employee

Accept or Kudo