4 Replies Latest reply on Jan 29, 2009 10:04 AM by SergeM

    super agent repositories guidance

      We have a large installation of epo 4.0 (10,000+ clients). I'm looking for guidance on where to put my super agent repositories. if you have x number of computers you have a repo, or x number of computers and the wan link is slower than y, you have a repo. I wish Mcafee had a document that gave this guidance.

      Thanks,
      Glenn
        • 1. RE: super agent repositories guidance
          Your superagents/repositories should be at least one on every subnet, preferably two according to McAFee. Which computers doesn't matter as long as they stay on and are connected. Could use local DCs or some other server/workstation.

          If you don't want to use that many atleast try to use on where the bandwidth is low, the agent will still communicate with the ePO but will get updates from Superagents/repositories. If you spann across continents consider using more than one ePO and use a master ePO to collect from the slave ePO/ePOs.
          • 2. RE: super agent repositories guidance
            SergeM


            I thought that was (more or less) what DR were : slave ePOs ?
            Or do you actually mean installing (separate) ePO Servers ? What would the advantage be in having several actual servers as opposed to one server and several DRs ?

            Serge
            • 3. RE: super agent repositories guidance
              Not sure what you mean by DR, but I think I understand what you mean...


              With super agents and repositories there are still traffic between the clients and the ePO server, setting up slave ePO/s will make the agents talk only to the nearest ePO thus eliminating the traffic to the master ePO (could use master ePO as fall back).

              I'm not saying that one of the options is better than the other but it's a possibility if you got bandwidth problems. You still get all the stats on the master ePO with that approach.
              • 4. RE: super agent repositories guidance
                SergeM


                DR = Distributed Repository

                I know a SuperAgent is one way of making a distributed repository and I know there are other ways. :cool:
                Creating a DR is one of the things on my plate, it's not top priority but close enough that it bothers me. I've been reading on it for some time and tried once ~2 weeks ago and it didn't work because I hadn't understood that with the SuperAgent (as it was set-up by default ?) the server is trying to copy files onto the superagent through MS netbios shares and it seems that our firewalls don't let this through "simply or easily". :eek:

                Now what I don't get is why one would set separate ePO servers if distributed repositories are good enough...

                I'm thinking about setting up DR in a few distant locations where the link is rather slow so as to minimize bandwidth usage. Installing a real ePO server would possibly work as well, but I'd have to find a way to remote admin the server since there is no real IT at those distant locations... :rolleyes:

                ta-ta-l8r
                Serge