Check first on the endpoint for the EDRDeployData folder and the logs contained within it. Those should help identify the problem, whatever it is.
In the (unlikely) event that it’s the same specific thing I mentioned about failed site gathers…on the client, check the directory “/var/opt/BESClient/__BESData/Patches for CentOS6 Plugin R2” (or the CentOS7, or whichever version you’re using. There should be a number of .fxf files in there, as well as a couple of binaries that are used during the prefetch phases - specifically “centos-client-x64” and “centos-client-x32”. If those are missing, we’d begin troubleshooting why the client is failing to gather those sites correctly, which may lead us up the relay stream to find a failing relay.
Actually this brings to mind another possible edge case. If your client is not subscribed to the “Patches for Centos X” site, but is subscribed to the site in which you have your baseline, you could encounter the same type of behavior. The Baseline is relevant, but the prefetch statements depend on the binaries in the Patches for Centos X site, so an action would present an “error” result when the prefetch plugin operations fail.