We have an issue w/ a cluster install on windows box where when we fail the cluster over to the other node, one of the agents installed on it gets hosed most of the time. It is a dual agent install of 1 prod & 1 dev agent - mostly the dev agent gets hosed so just uninstall/reinstall but had issues w/ prod & couldn't get it to fully clear this time so had to do work around to get jobs running again. Trying to figure out fix. I install the agents as local install on cluster local node on the cluster drive - one for prod & different one for dev. Should I use the cluster group. Install on cluster virtual server group instead? Person I took over for said to do it as first way I had been doing but wondering if it should be the other to prevent issue I'm seeing. When issue occurs, when looking at what active node should be, I only see one agent using the ctmagcfg or ag_diag_comm commands instead of the 2 I should see since it's a dual agent install but when I look at the Program Features both agents show as installed plus when I go to the directory of where "missing" agent is installed, the folder & appropriate stuff is there.