Kudos goes out to the Accelerite team for quickly addressing an issue we reported. I have been a proponent for creating a less RCS intensive Patch Manager product for quite a while. When the "Patch with Metadata Only" option came out I believed this was the answer. It was actually called "Offline Scanning" or "OPUS" (no idea what that stood for) at certain points. That model downloads all the patch data needed to do discover_patch to the device so it does not have to talk to the RCS during the scanning. I had assumed that meant it would disconnect from the RCS and not be using an Active Task. We had designed our capacity around this fact. However, after troubeshooting our infrastructure issues during our production patch rollouts the last couple of months I discovered it was not working that way. It was not actually talking to the RCS, but it was still holding an Active Task and connection to the RCS during the 3~5 minutes or so it took to do discover_patch (all locally). I am referring to the Microsoft patches BTW.
Accelerite understood the issue, addressed it, and now it is already fix for the next time you do a patch acquire. I was impressed.
Long time customers may remember that a big selling point of Radia over EDM when it was first released was that EDM held the connection to the RCS (manager) during the entire connect. With Radia is was broken up so the client would connect and disconnect multiple times as to not waste resources on the server side while the client was busy doing stuff locally. We need to make sure we maintain that model. It helps us customers keep our cost down in infrastructure. I actually have an enhancement request for disconnecting during the BDELETE for similar reasons.
I have not implemented in Prod yet (it will not be until Jan patches are released), but I will repost any interesting results in reduce capacity needs after I analyze them.
