Forum Discussion

Douglas_Gaigher's avatar
10 years ago
Solved

Move Partitions to a new Storage with native deduplication

Hello,

The scenario is a EV Server with about 15 TB of data with OSIS disabled and NTFS LUNs used as VS Partitions. A new storage with deduplication capabilities was aquired.

We would like to know if it is possible to move the partitions to new LUNs located at the new storage and make benefit of it's deduplication capability to reduce the volume of data.

Is there any restriction on using the storage-level deduplication? it must be supported by Veritas? It is just a matter of a normal move partition operation?

Thanks

  • @Doug, Yes, you can enable VNX Block deduplication. 

    Remember it's fixed block dedupe within a pool of thin LUNs and not global across the storage array.

    Downside - It will yield less storage effiencies than EV SIS might have done.

    Upside - You don't have to do any complex migrations to achieve the savings. Just relocate the partitions in the standard way.

  • if every vault store partition has always had the options checked for "device performs data deduplication" and "device performs data compression" then yes. also, according to the latest compatability charts:

    Storage devices: CIFS/SMB

    Enterprise Vault support for storage devices using
    access over the CIFS/SMB protocol. For EV 10 and 11.

    EMC VNX Series (DART 7.0.12 or later) 
    EMC VNX Series (DART 7.1.47.5 or later)

     

  • This means that it is not supported to use NTFS volumes to store partitions on this device?

  • If you want to just use them as a NTFS share and access via a UNC you will be fine but you will not be able to use the device built in de-duplication unless it utilizes the same DART version in the compatibility doc.  

    If you are going to use the partitions built-in deduplication I am pretty sure you will need to either use Move Archive and a third-party app like Archive Shuttle to migrate the archives to the new Vault Store using that device as partitions so the data can run through the API again.

     

  • if you're presenting this via CIFS/SMB then it's just NAS and not block or object which means you can just copy the old partitions over to the new storage and update the vault store partition paths in the database.

  • @Doug,

    Will you be using the EMC VNX as block-level storage (LUNs) or file-level storage (CIFS) ?

    Do you want to enable VNX deduplication on LUNs or Enterprise Vault SIS on CIFS vault stores?

    I think you need to internally clarify what you objectives are.

    Finally what sort of storage are the NTFS LUNs on currently ?

    D
     

  • Hello guys and thanks for the answers. I will try to clarify my question:

    Actually Block-Level storage (LUNs) formated as NTFS are presented to the EV server and used to store the VS Partitions. The actual Storage is the IBM XIV. OSIS is not enabled, so there is duplicated data on the Vault Store.

    We want to migrate the Vault Store Partitons to new Block-Level storage (LUNs) formated as NTFS, provided by the EMC VNX 5200. With this configuration is possible to make benefit of the Storage-Level deduplication and compression?

     

    Thanks

     

     

  • in that case EV will be configured the same way you have it now, with NTFS type partitions so you can just copy the old partitions over to the new storage and update the vault store partition paths in the database.
  • @Doug, Yes, you can enable VNX Block deduplication. 

    Remember it's fixed block dedupe within a pool of thin LUNs and not global across the storage array.

    Downside - It will yield less storage effiencies than EV SIS might have done.

    Upside - You don't have to do any complex migrations to achieve the savings. Just relocate the partitions in the standard way.