locked
Backup Performance (is this behavior normal?) RRS feed

  • Question

  • Hi, another newbie performace question.  I've done my initial backups from various PC just fine.  I added another 1TB had to my WHS, added it to the Storage Pool and then from my HTPC added one of it's 1TB Media drives to back-up.  The problem it is soooo slow I've been doing some testing to see where the bottle neck is.

    The following are the average values from observing (resource monitor, Jperf, and the Disk Mgt Plug-in for WHS) over 4 hours of what I think are the relevant processes from the PC in questions over the last few hours.  To me it lookes like backupengine is flogging the HDD creating a ShadowCopy but little is being transmitted to the WHS.

     kb/sec   Tested from the PC   Description 
          390,625  Jperf - TCP test   WHS ("SERVER") to PC 
          325,521  bakupengine.exe   \Device\HarddiskVolumeShadowCopy27 
               1,953  bakupengine.exe   Network "sending" to SERVER 
                   160  Disk Management Plug In   Disk 0 - in WHS 
                      -    Disk Management Plug In   Disk 1 - in WHS 



    Then all of a sudden the following started to happen (seems the data is now being sent as I would have expected):

     kb/sec   Tested from the PC   Description 
          156,250  bakupengine.exe   \Device\HarddiskVolumeShadowCopy27 
          156,250  bakupengine.exe   Network "sending" to SERVER 
                   128  Disk Management Plug In   Disk 0 - in WHS 
          152,000  Disk Management Plug In   Disk 1 - in WHS 


    Is this normal "behavior" or have I got something stuffed up?  eg I'm thinking it:
    - First 4 Hours, calculating and sending the Sector Hash to the WHS
    - from then on sending the actual sectors.

    Thanks
    Nathan
    • Edited by jmone Sunday, March 1, 2009 6:20 AM
    Sunday, March 1, 2009 2:35 AM

Answers

  • That's fairly normal behavior, I think. Is it the first backup for that computer? If so, the length of time it took would be fairly normal as well; Windows Home Server needs to get all the clusters (or at least their hashes) into the backup database.

    The backup engine uses volume shadow copy services to get a consistent "point in time" snapshot of a disk, so anything that happens on the client PC between when the backup is initiated and when it ends doesn't result in an internally inconsistent backup. So you will see VSS activity during that period.

    I'm not on the WHS team, I just post a lot. :)
    • Marked as answer by jmone Monday, March 2, 2009 11:20 AM
    Sunday, March 1, 2009 6:07 AM
    Moderator
  •  Hi Ken. 

    It is hopefully the first backup of this drive (still going)...as the last two attempts failed at the same point so I've since installed a Hot Fix - fingers crossed.  To clarify my post:
    1) the first 4-hours (9am-1pm) was the first set of logs where little was transferred to the WHS server itself but backupengine was very busy
    2) the second set of logs is typical throughput I've been seeing over the last 4-hours+ (1pm - now) when there was a big jump in the volume of data being sent to the WHS servers.  I calculate at this rate I've still got hours to go.

    Due to the failed attempts and the time it takes for the initial backup I started watching what was happening and wanted to check that the behavior I was seeing was expected.

    Thanks
    Nathan
    • Marked as answer by jmone Monday, March 2, 2009 11:20 AM
    Sunday, March 1, 2009 6:33 AM

All replies

  • That's fairly normal behavior, I think. Is it the first backup for that computer? If so, the length of time it took would be fairly normal as well; Windows Home Server needs to get all the clusters (or at least their hashes) into the backup database.

    The backup engine uses volume shadow copy services to get a consistent "point in time" snapshot of a disk, so anything that happens on the client PC between when the backup is initiated and when it ends doesn't result in an internally inconsistent backup. So you will see VSS activity during that period.

    I'm not on the WHS team, I just post a lot. :)
    • Marked as answer by jmone Monday, March 2, 2009 11:20 AM
    Sunday, March 1, 2009 6:07 AM
    Moderator
  •  Hi Ken. 

    It is hopefully the first backup of this drive (still going)...as the last two attempts failed at the same point so I've since installed a Hot Fix - fingers crossed.  To clarify my post:
    1) the first 4-hours (9am-1pm) was the first set of logs where little was transferred to the WHS server itself but backupengine was very busy
    2) the second set of logs is typical throughput I've been seeing over the last 4-hours+ (1pm - now) when there was a big jump in the volume of data being sent to the WHS servers.  I calculate at this rate I've still got hours to go.

    Due to the failed attempts and the time it takes for the initial backup I started watching what was happening and wanted to check that the behavior I was seeing was expected.

    Thanks
    Nathan
    • Marked as answer by jmone Monday, March 2, 2009 11:20 AM
    Sunday, March 1, 2009 6:33 AM