3 Replies Latest reply on Dec 16, 2014 1:26 PM by JoeBidgood

    McAfee Agent 5.0 slow replication


      Hello, since upgrading the McAfee Agent to 5.0 on the SuperAgent distributed repositories, the incremental updates are really slow.

      I am checking the communication between the ePO 5.1.1 and the repository with TCPView and the Resource Monitor and it appears that a full replication is actually occurring, rather than just incremental.

      Also the speed is rather slow - around 350 kB/s and takes around 40 minutes to complete for 750 MB of data.

      Changing any option in the McAfee Agent General policy, including disabling of "Run agent processes at lower CPU priority" does not make any difference. Rebuilt the repositories from scratch - the same issue occurs, the incremental replication takes as much time as the full one.

      Upgraded the extension from version to - no difference.


      If I return the old agent 4.8 on the repository the full replication runs with up to 5-7 MB/s and finishes for 5 minutes, the incremental finishes in 10-20 seconds without any traffic.


      So there seems to be two issues with the new Agent 5.0 - low bandwidth utilization for the replication and disregarding incremental replication and performing full one instead.


      Has anyone noticed the same?

        • 1. Re: McAfee Agent 5.0 slow replication

          Would you do me a favour - could you open a case for this with Support, and ask for it to be immediately escalated to me?


          Thanks -



          • 2. Re: McAfee Agent 5.0 slow replication

            We upgrade our remote repositories (super agents) to version 5.0 on the 25th of Nov. The very next replication we noticed it takes 6-7 hours for an incremental replication, vs under 10 minutes for 27 remote repositories.  Support wants me to downgrade all the agents to version 5.  Anyone else seeing this issue?

            • 3. Re: McAfee Agent 5.0 slow replication

              An issue has been identified with superagent replication on MA5 agents. The developers are working on a fix for this, but in the meantime the recommendation would either be:


              a) Use lazy caching, instead of replication tasks, which will probably also save you some bandwidth, or

              b) downgrade the superagent machines to the latest version of MA 4.8 (currently 4.8 Patch 2).


              There's no need to downgrade *all* the MA5 clients, as this only affects superagent replication. I'll create a KB for this.


              Thanks -