Convert TPVV to Dedup

[rashminder]

Hi Experts,

I'm converting Thin LUN to a Thinly dedupe LUN using tunevv command. Will this cause data to be lost?

Command: tunevv usr_cpg CPG_DD -tdvv SAN_TN

...Please suggest...

[Sheldon Smith]

No.

What would make you think otherwise?

[rashminder]

I have converted a test LUN which was 17% full as per IMC, but after running this command, its 0%

The LUN was not in use, so, I don't know about impact (if any).

I'll be freeing up couple of LUNs tomorrow & test before running on production data.

[Dennis Handly]

>I have converted a test VV which was 17% full as per IMC

What does the CLI say about usage?  Before and after?

[Gary_L]

Keep in mind that the view on a VV space when it's a TDVV vs a TPVV are different.  Becasue we have data in the DDS and the client volume or DDC we show more of a load factored view of the space so you have to be careful in identifying capacity on a per volume basis.

For instance, if you had 3 volumes that were identical deduplicatable data and they were each 10G of space on a 100G volume.  That 10G would be in the DDS and the space for each volume would show aproximately 3.33G each because they are all contributing to the DDS and the load is distributed for the data shared.  If you removed one of thoes volumes, the space in the DDS doesn't change, (still 10G) but the two volumes would now report 5G each.

You will not lose data on any conversion, that should not be a worry.

Gary

[sunshuai]

Can you explain DDC and DDS in detail? DDS is shared with the same VV in CPG, is DDC present in every VV?

[Gary_L]

Correct, the DDS is the shared space for all of the TDVVs in the same COG. DDS can grow up to 64TB.

DDC is just the dedup client volume or the actual VV you created. The DDC holds data as well.

Gary

[sunshuai]

Ok, thanks, I understand that. I have another question.

After enable deduplication and compression.Why all writes with new hashes are written to the private space（DDC） first,not shared space（DDS）,when data is seen for second time ,the data is moved from private space(DDC) to shared space(DDS).This will not consume resource？Why cant write DDS in the beginning？Why keep such a big DDC and so small DDS？

[sunshuai]

In the past, only deduplication, DDC space is particularly small, only a very small hash of conflict data is stored, why later joined the compression function to make such a big change, a very large DDC and small DDS .

[sunshuai]

I have seen a document before saying that there will be an Express Scan before doing the compression. What does this Express Scan use to determine if this part of the data is compressible?

[Gary_L]

sunshuai,

Express Scan is used to prevent wasted CPU cycles, if a page is not compressible then it is written to disk uncompressed.  This is instead of compressing the entire page only to find out the data was not compressible and using resources to gain limited/no return.

[sunshuai]

Thank you Gary, you answered my confusion well.