DPM 1807 (and 2016 for that matter) has very little network I/O when I would expect it to be very high RRS feed

  • Question

  • Part of my many problems that I have had since DPM 2016 (and now 1807) has been released is that I have many jobs queued and running but I observe very little network I/O.  We're talking in the Kbps range, not the Mbps range. The system has plenty of RAM, CPU power (12 physical cores/24 logical). The hardware seems to be doing nothing at all but there are all these jobs queued. Any one else observing this too?  No network throttling is turned on.  There is nothing wrong with the network.  No other systems are seeing this issue--only DPM.

    When I used DPM 2012 R2 (and earlier), the network I/O was always extremely high (as expected) when these jobs were queued.  The jobs would consume the bandwidth and then finish.  With 1807/2016, the jobs just seem to sit in the queue and take forever.  Occasionally, I see high spikes of activity, but this rare when I would expect it to be common with the jobs queued.

    Thursday, February 21, 2019 3:05 PM

All replies

  • Hello!

    Do you see anything out of the ordinary in the DPM logs?

    What kind of backup are we talking about, disk, tape or online backup?

    Could you also tell us what kinds of workloads are you backing up with DPM? For example:

    • Windows Servers
    • SQL Server databases
    • SharePoint
    • Exchange
    • Hyper-V virtual machines
    • File Server

    Best regards,

    Blog: LinkedIn:

    Thursday, February 21, 2019 3:42 PM
  • Nothing out of the ordinary in the logs.  I'm backing up all of those workloads, but I was also doing that with DPM 2012 R2.  I'm going to trying some swapping around some hardware just to make sure nothing is a problem there.
    Thursday, February 21, 2019 4:21 PM
  • Okay, let us know how it goes!

    Blog: LinkedIn:

    Thursday, February 21, 2019 11:18 PM
  • Hi,

    Just checking to see if you have any update?

    Blog: LinkedIn:

    Friday, March 8, 2019 12:05 AM
  • Possibly.  I'm working with a MS support engineer through a premier case on this issue (as I have been on and off the for last 2+ yrs).  We reinstalled the primary DPM server but this time instead of using Windows Storage Pools we used the native RAID6 on the controller.  That's what I always had done before DPM 2016 and MBS, but not since then.  So far, it seems to be a major improvement although it's only been running a couple days.  I need more time to be certain.  I'm going to also be setting up my secondary in the next few days with this same configuration.  If that is fast as well, that may be the cause for us as both the primary and secondary were slow (but they both have essentially the same hardware).

    Friday, March 8, 2019 2:28 PM
  • Okay, thanks for sharing this information, let's hope everything turns out well for you!

    Blog: LinkedIn:

    Friday, March 8, 2019 2:46 PM