none
DPM2007SP1, Slow Sync RRS feed

  • Question

  • I have been running DPM 2007 for several years now.  Minor problems, but all-in-all, its been great, still is -- I just now have my first problem, =)

    I have 8 sites that I back up to a central site, all across bonded-T1s.   Some sites have only 1-2 servers, other have 8-10. 

    As of about 2 weeks ago, two specific servers have begun to synchronize very slowly.  Not that I think its relevant, but both are in sites with 2 systems.

    Anyway, two of the servers from two seperate sites are now synchronizing increadibly slow.  It is acting as if, from my perspective, like the server is throttled.

    For example:  In Site A:  One server syncs about 3GB, the second server in the same time will sync about 400-500MB.  The same is essentially true at site 2. 

    Event Logs on sending and receiving servers are clean.

    Ample unused bandwidth at both sites.

    Disk Channel, NIC, etc all test fine on all ends.  Counters are way below any problem status.

    As a test, from both 'slow' DPM servers, I tested standard file transfer rates.  The same server that was only able to DPM 300MB in the same time it transferred 2.8GB, from same server to same server.  DCOM looks clean, window sizing looks clean.

    I understand that 'performance problems' are incredibly difficult to answer/help on a forum, but I am honestly out of ideas.  Hoping that perhaps someone out there has seen something similar or had an experience that can help.

    One thing I have not done is delete and recreate the PG.  With over 500GB a piece, thats a lot of effort to undertake when its possible slow link issues may further complicate synchronizations.

    Thanks to all

    Friday, April 29, 2011 7:22 AM

All replies

  • Hello,

    Thank you for recognizing that troubleshooting performance issues via forum can be difficult. 
    I do have some questions for you though to help narrow down the possible cause.

    Questions:
    a.) If you take a netmon trace do you see a lot of retransmitts? a lot of resets?
    b.) Do you have chimney or rss enabled?  netsh int tcp show global
    c.) What are the datasources and what are there sizes?
    d.)Is this for all datasoruces or just specific ones?
    e.)You stated that the counters are way below any problem status, did you also include "Output Queue Length" for the NIC?

    Output Queue Length description: 
    Output Queue Length is the length of the output packet queue (in packets). If this is longer than two, there are delays and the bottleneck should be found and eliminated, if possible. Since the requests are queued by the Network Driver Interface Specification (NDIS) in this implementation, this will always be 0.

    f.) Event logs on both sides do not have any errors or warnings in regards to "delayed write failed", vss related or loss of communication?
    g.) are you using compression on the wire?
    h.) Is this link encrypted or compressed via thrd party means?
    i.) You stated "I tested standard file transfer rates." how? Copy and paste? Third party util? Did you test with the same datasource that DPM was backing up?
    j.) What's the amount of RAM installed on the DPM and the protected server?
    k.) What is the Page File size on the DPM and File server?


    DPM performance articles:
    DPM Managing Performance: http://technet.microsoft.com/en-us/library/ff399471.aspx
    How DPM Operations Affect Performance: http://technet.microsoft.com/en-us/library/ff399315.aspx
    DPM and Memory: http://technet.microsoft.com/en-us/library/ff399244.aspx
    Performance Counters: http://technet.microsoft.com/en-us/library/ff399366.aspx
    Improving Performance: http://technet.microsoft.com/en-us/library/ff399571.aspx
    Managing DPM Performance on a WAN: http://technet.microsoft.com/en-us/library/ff634226.aspx
    How Protection Groups Changes Affect Jobs: http://technet.microsoft.com/en-us/library/ff399316.aspx

     

    Thanks
    Shane

     

    Friday, April 29, 2011 12:40 PM