none
Move Disk Storage slowtransfer Speeds Version 1807 RRS feed

  • Question

  • Attached new storage pool volume on from seperate piece of hardware. Went to Move disk storage and it started out "ok" probably transferring about 1GB/hour and now its hugely pathetic at 40MB/hour.

    What's the deal?

    How does one get this to transfer at acceptable rates?

    Saturday, October 19, 2019 8:54 PM

All replies

  • Hi,

    There's been many cases with slowness when performing the "Move Disk Storage" for unknown reasons, but you're not alone, see DPM 2016 Move Protection Groups / Datasources

    I know this isn't something you want to hear, but unfortunately I don't believe there's anything else you can do other than wait until the move is completed, 

    Might want to make sure there are no backup jobs or anything else pressurizing the DPM server while performing the "Move Disk Storage".

    You may also try using the PowerShell method:
    SC 2016 DPM UR4: Migrate Backup Storage in 3 simple steps

    Best regards,
    Leon


    Blog: https://thesystemcenterblog.com LinkedIn:

    Saturday, October 19, 2019 9:20 PM
  • Thank you.

    Is there a way to stop the migration process? Its making all my other jobs slow now.

    Saturday, October 19, 2019 10:22 PM
  • I don't believe you can stop it, and if you could it could cause corruption.

    How much data are we talking about?


    Blog: https://thesystemcenterblog.com LinkedIn:

    Saturday, October 19, 2019 10:24 PM
  • about 500GB. in 6 hours its only transferred 7 GB.
    Saturday, October 19, 2019 10:32 PM
  • Okay, there may be many reasons to why it can be this slow, slow disks, not enough resources on the DPM server, too much activity (DPM jobs).

    If you started the "Move Disk Storage" from the DPM console, you will not be able to cancel it, unless you force it which I strongly wouldn't recommend, as it can lead to you losing your data or corrupting the data.

    I would make sure that there are as few backup jobs or activity as possible, preferably none on the DPM server while doing this move of data.



    Blog: https://thesystemcenterblog.com LinkedIn:

    Saturday, October 19, 2019 10:45 PM
  • Thank you for the feedback.

    No that's the sad part. That's the speed I got after I cancelled all jobs. So there was no strain on the server whatsover. The storage is connected via 10G fibre and its literally 1 foot away from the dpm server.

    The server is a beast. 330GB of RAM and its not even using 10% of the or CPU.

    The reason for migrating was because I'm suffering from the infamous ReFS slow down which currently has no known solution. I really didn't want to lose all my existing backups on disk so I figured I would try to stick it to the man by migrating everything off the storage pool whose backups get slower by the weeks and migrate it onto a freshly formatted ReFS storage pool.

    But I guess the joke was on me. Migrations are even worse!

    Interestingly enough after I started a a job the speed started to pick up....so itsead of 50MB every 30 minutes it went to 50MB every 5 minutes.

    Saturday, October 19, 2019 10:52 PM
  • I've witnessed this process can take days, but I'm not aware of any method to speed this up.

    How does it look for you now?


    Blog: https://thesystemcenterblog.com LinkedIn:

    Sunday, October 20, 2019 3:02 PM
  • Highly erratic.

    Absolutely 0 stress on the server = 40MB/30 minutes

    Running only Tape Backups = Up to 13.5GB/Hour

    Running Recovery points and Tape Backups = Slowly drops from 13.5GB/Hour to around 5GB/Hour.

    I feel like placing a call to Microsoft so they can troubleshoot but something tells me that I'm going to end up in the discarded pile of unsolved tickets.

    I have an identical setup at another location which presents 0 issues. The only difference is that that setup only backs up under under 15TB and this location backs up about 20TB.

    Is there a cut off point where it stops working ?

    Monday, October 21, 2019 1:18 AM
  • Is the 20TB on one or more LUNs?

    I assume it would stop if the disk or the DPM server becomes unavailable.


    Blog: https://thesystemcenterblog.com LinkedIn:

    Monday, October 21, 2019 5:52 AM
  • one LUN.
    Monday, October 21, 2019 2:38 PM
  • Okay, has this DPM environment been upgraded from an older version than 1801?

    Blog: https://thesystemcenterblog.com LinkedIn:

    Monday, October 21, 2019 2:46 PM
  • no. It was completely wiped out. Fresh OS install 2016 Standard and Fresh installation of 1807.
    Monday, October 21, 2019 5:25 PM
  • Then that should be just fine, I can't come up with anything that would cause this kind of slowness, however if this is causing problems to your production environment you should consider raising a ticket to Microsoft.

    https://support.microsoft.com/en-us/hub/4343728/support-for-business


    Blog: https://thesystemcenterblog.com LinkedIn:

    Monday, October 21, 2019 6:07 PM
  • Hi,

    Just wanted to check if you have any update regarding your issue?


    Blog: https://thesystemcenterblog.com LinkedIn:

    Wednesday, November 6, 2019 9:09 PM