I have a pretty specific requirement, detailed below. Here's what I'm not looking for help for, to keep things tidy and on topic:
- How to configure a standard proxy
- Any ESX setup required to facilitate traffic sniffing
- How to sniff traffic
- Any changes in design (my scope limits me)
I need to setup a test environment for a network-sniffing based HTTP app monitoring tool, and I need to troubleshoot a client issue but he only has a prod network, so making changes to the config on client's system "just to try" is costly.
The goal here is to create a similar system in my lab, and hit the client's webapp and redirect my traffic - using a proxy - into the lab environment. The reason I want to use a proxy is so that only this specific traffic is redirected for all to see, and not all my web traffic (like my visits to serverfault :P).
Everything will run inside an ESX 4.1 machine.
In there, there is a traffic collection vSwitch in promiscuous mode that is not on the local network for security reasons. The VM containing our listening agent is connected to this vSwitch.
On the same ESX host, I will setup a basic linux server and install a proxy (either apache + mod_proxy or squid, doesn't matter). I'm looking for ideas on how to deploy this for my needs so I can then figure out how to set it up accordingly.
Some ideas I've had were to setup two proxies, and have them talk to each other through this vSwitch in promiscuous mode, but it seems like a lot of work. Another idea is a dual-homed proxy, but I've never seen/done that before so I'm not sure how doable it is for what I'd like.
I am OK with setting up a second vSwitch in promiscuous mode to facilitate this if need be, but I cannot put the vSwitch on the lan (which is used so my browser would communicate with the proxy) in promiscuous mode.