none
MBS volume > 100TB? RRS feed

  • Question

  • Hello, we have the following config:

    - physical server with mix of SSDs and HDDs, which are used in a storage spaces pool and there are 2 ~60TB tiered volumes/disks created

    - these 2 60TB disk are used as a storage for 1TB VHDX files for DPM VM (running on the physical host), and inside DPM these 1TB files are used to create simple storage space, which is in turn added to DPM as MBS storage

    The 2 tiered volumes on the host are ~60TB, because we want to use deduplication, which is limited to 64TB volumes.

    Now we have 2 options for the DPM storage:

    1) create ~60 (as a start before taking dedup in consideration) 1TB VHDX files per tiered volume (so 120 total) and create 2 simple storage spaces inside DPM, honoring the 'physical' tiered volume layout on the host

    2) create ~60 (as a start before taking dedup in consideration) 1TB VHDX files per tiered volume (so 120 total), but create only 1 simple storage space inside the DPM spanning over both tiered volumes on the host

    I'd like to go with scenario 2 since it provides more flexibility (don't have to deal with 2 volumes inside DPM - choosing which protected datasource is located on which volume, moving backup data when one volume becomes full etc.) and also better cache (mirror tier) utilization on the host. I haven't found any limits preventing this (120 1TB VHDX making up single 120TB storage space), but I'd like a confirmation/opinions.

    Thanks in advance

    Wednesday, February 6, 2019 10:03 AM

Answers

  • Hello!

    In your case I would also go with option 2, from DPM 2016 and newer versions there are no longer any Local Disk Manager (LDM) limits or storage over-allocations.

    So you should be good to go!

    Best regards,
    Leon


    Blog: https://thesystemcenterblog.com LinkedIn:

    • Proposed as answer by Leon Laude Wednesday, February 6, 2019 4:51 PM
    • Marked as answer by MarkosP Monday, February 18, 2019 8:20 AM
    Wednesday, February 6, 2019 10:38 AM

All replies

  • Hello!

    In your case I would also go with option 2, from DPM 2016 and newer versions there are no longer any Local Disk Manager (LDM) limits or storage over-allocations.

    So you should be good to go!

    Best regards,
    Leon


    Blog: https://thesystemcenterblog.com LinkedIn:

    • Proposed as answer by Leon Laude Wednesday, February 6, 2019 4:51 PM
    • Marked as answer by MarkosP Monday, February 18, 2019 8:20 AM
    Wednesday, February 6, 2019 10:38 AM
  • Thanks for your input Leon.

    I went on with option 2 and all went well. Actually, I already had DPM MBS using 56 VHDX from one of the tiered volumes, so I was in fact adding another 56 from the 2nd tiered vol. Only resizing the partition took rather long (about 2mins).

    Wednesday, February 6, 2019 12:56 PM
  • Yes the resizing might take a bit of time the bigger they get, but it's nothing to worry about, just have patience.

    Blog: https://thesystemcenterblog.com LinkedIn:

    Wednesday, February 6, 2019 1:01 PM
  • Hello MarkosP,

    Just checking in to see if you have any update?


    Blog: https://thesystemcenterblog.com LinkedIn:

    Friday, February 15, 2019 9:04 AM
  • Well not really, it worked as I've wrote before. One thing I didn't realize/know was that when I was expanded the existing pool, the newly added VHDXs (from the 2nd tiered volume) are not actually being utilized - I guess Windows added the capacity to the end of the volume when expanding.

    I've tried running Optimize-StoragePool from inside the DPM, but it finished in couple seconds and seemed to have no effect on the data distribution.

    Monday, February 18, 2019 8:20 AM