Skip navigation
McAfee Secure sites help keep you safe from identity theft, credit card fraud, spyware, spam, viruses and online scams
This discussion is archived
876 Views 7 Replies Latest reply: Dec 11, 2012 9:54 AM by JoeBidgood RSS
tbujnowski Newcomer 36 posts since
Jul 16, 2011
Currently Being Moderated

Dec 6, 2012 10:12 PM

Replication Tasks in parallel

Is there a way to raise the number of replications task in parallel?

 

When issue a replicate now command, no matter how many repositories I select, it only replicates six at a time.

 

I found a document on the net that says that its replications concurrence is equal to number os CPUs x 6. And that matches my scenario.

 

Is there a way to raise this number? (Besides raising the number os CPUs)

 

Thanks in Advance!

  • JoeBidgood McAfee SME 2,880 posts since
    Sep 11, 2009
    Currently Being Moderated
    2. Dec 7, 2012 10:14 AM (in response to tbujnowski)
    Re: Replication Tasks in parallel

    This is still valid, but I'd suggest experimenting with caution as setting this too high can have a performance impact on ePO when replications are running.

    Out of interest how many DRs are we talking about?

     

    HTH -

     

    Joe




    (Please post questions to the forum, as I am unable to respond to private messages. Thanks!)



  • DonHanson McAfee Employee 7 posts since
    Aug 16, 2010
    Currently Being Moderated
    3. Dec 7, 2012 5:52 PM (in response to JoeBidgood)
    Re: Replication Tasks in parallel

    Have you looked at using Super Agent repositories with Lazy Caching to see if they would fit your environment and needs?  

     

    Lazy caching turns the repository replication model from a push to a pull model.  It can save significant bandwidth transmission as only the files necessary to update your machines are pulled across the wire.  The main consideration is that updating of DR is no longer scheduled, so it can happen at any time of the day as machines look for updates.

  • JoeBidgood McAfee SME 2,880 posts since
    Sep 11, 2009
    Currently Being Moderated
    5. Dec 10, 2012 11:46 AM (in response to tbujnowski)
    Re: Replication Tasks in parallel

    I think the files are cached for 30 days, by default, so they will be removed 30 days after the last time a client requests them.

    In the scenario you describe I definitely think you should be looking at lazy caching - replicating over 100MB of DAT file every day is not going to be an efficient use of that bandwidth, I feel.

     

    HTH -

     

    Joe




    (Please post questions to the forum, as I am unable to respond to private messages. Thanks!)



  • JoeBidgood McAfee SME 2,880 posts since
    Sep 11, 2009
    Currently Being Moderated
    7. Dec 11, 2012 9:54 AM (in response to tbujnowski)
    Re: Replication Tasks in parallel

    Whatever file is requested will be downloaded - so if all the machines downstream of the SA repo are only one day out of date, then all that will be downloaded is one incremental file.

    However if a new machine is brought online, then it will request the full DAT.

     

    HTH -

     

    Joe




    (Please post questions to the forum, as I am unable to respond to private messages. Thanks!)



More Like This

  • Retrieving data ...

Bookmarked By (0)

Legend

  • Correct Answers - 5 points
  • Helpful Answers - 3 points