5 Replies Latest reply on Aug 19, 2008 12:22 PM by SergeM

    Distributed Repository

    SergeM
      Hi,

      I've been managing our ePO servers for about 3 months now, so I'm quite new at this.
      We presently have two different servers (S1 = ePO3.6 and S2 = ePO4.0 ; the 3.6 will be upgraded in the near future).
      For geographic reasons, we need to install Distributed Repositories (DR) in some distant sites.

      I've done some reading (ePO4 Install Guide & Product Guide) but am not 100% sure of how this works.


      1. From the manuals it somehow looks like a DR is not much more than a PC on which I'd promote the McAfee Agent to SuperAgent - is it that simple ? In another place it looks like a full server install ? :confused:

      2. How do I install/create a DR ? I've read the manual but I'm not sure how to proceed.
        Do I take a server (PC with SQL DB) and then install an ePO Server ?
        How do I then declare it as a DR to the first SVR ?
        Do the DR pull the updates/packages from the main SVR or do they connect to the nai/mcafee servers ?

      3. Bonus : could I use one of the two servers I already have as a fallback site in case the other one can't connect to McAfee FTP site for updates ? Is this recommended ?



      thanks for your answers
      Serge
        • 1. RE: Distributed Repository
          I am using super agents as distributed repositories in our locations worldwide

          and yes, it is that easy.

          just change the CMA / MA policy for that PC (I am using servers, because the repository should be available, you can also use a PC which is turned on all the time):

          1.) convert agent to superagent
          2.) wait until it is recognized as a superagent
          3.) enable the option that it should be used as a distributed repository and enter the path where you would like the files to be stored
          3.) replicate the master to the distributed repository

          what I am also doing is: I configure the agents just to use the repository at their location for the update, but you can leave all repositories in the agent-configuration and let him decide by ping (latency) which repository they should use
          • 2. RE: Distributed Repository
            SergeM


            So the repositories don't need to have an SQL DB ? Just enough disk space (and bandwidth) ?

            Considering the bandwidth being the problem (bottleneck) between the main server and the remote locations, how do you cope with uploading updates or new software (several MB worth of data) through a 512K line ? Even if we do this during the night (worldwide night?)

            Serge
            • 3. RE: Distributed Repository
              When i was setting up our EPO 4.0 config and our repos (we have about 240). Im sure the Tier 3 engineer said the only differance in converting them to SA was the encryption it uses.
              • 4. RE: Distributed Repository
                a superagent repository just needs disk space and bandwidth, thats correct

                do the locations have their own internet-access?
                if yes, you should think about using "McAfee AutoUpdate Architect" to create distributed repositories
                you should be able to add these to your ePO configuration.

                but I am not sure how to do all of this, since we provide internet access to our worldwide locations only from our headquarters.
                we are also using some optimizers to reduce the traffic and in the worst case we could also do it without optimizers, because we have enough bandwidth
                • 5. RE: Distributed Repository
                  SergeM



                  Thanks for (all) your answer(s)

                  Like you, we provide internet access to our worldwide locations only from our headquarters (rather we centralise access for some reasons). General bandwidth isn't normally a problem, just for a few locations with 512K lines.

                  thanks