I've got two, I guess pretty easy to answer, questions.
1. McAfee Agent Installation is not going trough any of the configured ports from ePO (Agent-Server-Communication - port 80, 443 etc.) Is that correct? We checked it with wireshark and we've seen it's using several ports like 137/445 etc.? Is there a way to install it from ePO from any of the configured ports on ePO?
2. Client Tasks / Product Deployment. We tested this as well on several subnets and it seems he always want to connect trough port 80 but instead we actually want ePO or the Client to use 443. Is this possible?
We are running ePO 4.5.937 and McAfee Agent 4.5.1270 on a Windows 2008 Server. Reason I'm asking this, is because we have a very restricted network and are trying to be as secure as possible.
Solved! Go to Solution.
1) Assuming you're talking about a push agent install, this is normal: the ePO server copies the installation package to the target machine using the standard windows admin$ share. If an agent that was being updated by a deployment task - i.e. one where the agent on the client is pulling the install package - the port used will depend on the settings of the repository being used (FTP, HTTP, etc.)
2) At the moment, assuming you're talking about pulling updates and deployment packages from the master repository, then I don't believe this is possible. If you want strict control over the port usage, you could create a distributed repository configured the way you want and configure the agents to update from there instead of from the master.
Thanks a lot for your fast answer!
1) solved then, thanks.
2) What exactly goes through the Server-to-Agent secure communication port then? Because what I see from the logs is when I try to deploy VSE 8.7i over ePO he's always trying port 80 which is the Server-to-Agent communication port and he doesn't seem to try the secure port at all?
Strictly speaking this is a bit simplistic. All agent-to-server communications (apart from items being pulled from the master repository, and wakeup calls) will take place over either the agent-to-server port, or the agent-to-server secure port, depending on which you have configured. This means all properties, tasks, events and everything else. There is no distinction between "confidential" information and anything else - it all uses whichever port you have configured.