none
DPM Transfer rate RRS feed

  • Question

  • we have 2 dpm 2012 R2 UR2 servers backing our 2010 SP3 DAG databases, one backs up the active DBs and the other the passive copies.

    I have noticed significant difference in the data throughput where the active copies get backed up in minutes ie. 1-2GB per minute whereas the passive copies get backed up at 100-200mb per minute.

    i know the active copies get backed up with the exchange store writer whereas the replication writer backs up the passive copies, is this by design?

    we are only taking a nightly sync once of the data

    currently the databases in total are <200GB but we are only part way through our migration so this will incrrease 3-4 fold in the next 6 months.

    is this the best way to back up the data in two locations without transfer across the WAN links

    Regards

    Paresh

    Saturday, June 14, 2014 1:01 PM

All replies

  • to add, replica creation is very fast, GB's per minute, just the nightly synchronization is extremely slow.

    Regards

    Paresh

    Saturday, June 14, 2014 4:40 PM
  • Active copies will be backed up by the Exchange Logs

    the passive Copie will be backed up the DB file, so thats the reason why passive backup is slower, cause DPM does now with the Log file where the different is since last backup, thats not the case if you to a passive backup


    Seidl Michael | http://www.techguy.at | twitter.com/techguyat | facebook.com/techguyat

    Saturday, June 14, 2014 6:09 PM
  • Is there any technical documents which states this?
    Saturday, June 14, 2014 7:25 PM
  • Seidi,

    Reading the blog article http://blogs.technet.com/b/exchange/archive/2012/07/09/everything-you-need-to-know-about-exchange-backups-part-3.aspx would go against what you stated above.

    In our scenario events 2021, 2110 and 2023 occur within seconds of the job starting, events 210/910 occur within seconds on the master db server.

    Event 963 occurs 25 minutes later to state the surrogate backup procedure to the master server has completed.

    the article state event 963 is logged when "When the files necessary for backing up DB1 are safely copied to backup media, the backup application signals VSS that the job is complete. VSS in turn signals the replication writer, and Exchange generates events 963 and 2046 on the passive copy server"

    so to me it appears the delay to actually getting the data to the backup server, a manual file copy between exchange and backup server takes 2 minutes to copy 20GB whereas the DPM is taking much longer to transfer the same data using a dedicated backup network.

    any help would be appreciated

    Regards

    Paresh

    Monday, June 16, 2014 1:45 PM
  • reviewing our logs, it appears the delay is at this stage

    Once the snapshots are created the backup application can copy blocks of data through VSS, which transfers blocks of data from shadow storage if they’ve been preserved due to a change, or from the actual disk volume if they haven’t. The replication service writer waits for the signal that the transfer of data is complete. This flow of data is represented by the purple arrows, which in this case indicates data getting copied out of the snapshots in storage, through I/O of the Exchange server, and on to the backup server.

    Can anyone tell how quickly the events 963,2046 on the passive node and 913/213 on the active nodes occur after the passive DB backup occurs within DPM.

    Thanks

    Paresh

    Monday, June 16, 2014 2:26 PM
  • Anyone from MS monitoring these posts?

    specifically our issue is with surrogate backups with DPM 2012 R2 UR2 backing up Exchange 2010 SP3 UR4.

    for some reason the time between events 960 and 963 there is a considerable time laspe.

     http://blogs.technet.com/b/exchange/archive/2012/07/09/everything-you-need-to-know-about-exchange-backups-part-3.aspx

    reading the above these events are where the copy of the snapshot data is taking place, we have a dedicated 1GB network between the dpm/exchange servers, but at during the duration of the copy of snapshot data the throughput is 10-15% whereas when we create a protection group to protect exchange data, this task occurs at a much higher throughput.

    our design is that DPM is taking backups of our passive nodes at our DR site. they are configured as full backup copy.

    anyone have any ideas?

    Thanks

    Paresh

    Friday, June 20, 2014 2:35 PM
  • Paresh,

    You are correct that the replication writer backups the passive copies.

    You stated there was considerable time lapse between 960 and 963 on the passive node.

    Using the document you listed when are you seeing events 2025 and 2027 and then 2029 and 2035 events after the 960 and before the 963 events?

    Taken from the article:

    "Events 910 and 210 on the active copy server, as well as 960 on the passive copy server, signify two things: First, they establish which server backing up a passive copy of the database; Second, the STORE service on the active copy server has marked the database with “backup in progress” in memory and acknowledges that the surrogate backup will proceed. Once this occurs it is not possible to backup the database again until either the current surrogate backup completes, or the “backup in progress” status is otherwise cleared.

    image

    Events 2025 and 2027 are generated when the replication writer prevents the replication service from writing logs copied from the active copy server to the local disk. Replay of logs also stops, thereby keeping the contents of the database files unchanged. At this point writes of data for the database getting backed up are “frozen”. VSS can now create the snapshots in shadow storage for each disk specified in the metadata.

    image

    VSS creates snapshots of disks D: and E:. Once these complete it signals the Exchange Writer, which in turns allows the replication service to resume log copy and replay. Events 2029 and 2035 are generated when the “thaw” is completed and normal disk writes are allowed to continue.

    image

    Once the snapshots are created the backup application can copy blocks of data through VSS, which transfers blocks of data from shadow storage if they’ve been preserved due to a change, or from the actual disk volume if they haven’t. The replication service writer waits for the signal that the transfer of data is complete. This flow of data is represented by the purple arrows, which in this case indicates data getting copied out of the snapshots in storage, through I/O of the Exchange server, and on to the backup server.

    image

    When the files necessary for backing up DB1 are safely copied to backup media, the backup application signals VSS that the job is complete. VSS in turn signals the replication writer, and Exchange generates events 963 and 2046 on the passive copy server. "

    ========

    If the 2035 comes quickly also on the passive node before the 963 there may be a transfer issue.

    So we have a few possible bottlenecks to look at.

    Disk I/O on both the DPM server and Exchange server. You can monitor resource monitor during a backup to check this.

    Network I\O issue - you stated that the passive is at a DR site so there may be more hops.  A network trace would help here.

    One test to potentially rule out disk on the Exchange node and network is to make a DB passive on the active node in the primary site and test a backup or vice versa.

    Andy


    Regards, A.Nadar, This posting is provided &quot;AS IS&quot; with no warranties, and confers no rights.

    Friday, September 5, 2014 4:07 PM
    Moderator
  • Any update?

    Regards, A.Nadar, This posting is provided &quot;AS IS&quot; with no warranties, and confers no rights.

    Tuesday, November 4, 2014 9:36 PM
    Moderator