We are using Windows Server 2022 RDS servers with RemoteApps and have just migrated to Kinetics 2024.2 and are still using the classic version. The development team is working toward migrating to the web version, but for now we are stuck using the classic/smart client. We always had various issues with running Epicor 10 on RDS servers, but they were pretty manageable and rare enough that it wasn’t a huge concern.
Some changes with either RDS 2022 or Kinetics 2024.2 or both have seemed to increase the sporadic issues we get with RDS servers. We are working with support and they have answered a lot of our questions, but I was curious to hear others’ experiences with running Kinetics on RDS servers and issues you may have encountered.
These are some of the biggest issues we have run into:
-
Groups of users will suddenly have very slow or unresponsive client activity, are kicked out, or are unable to get in at all with no error, while the server performance looks normal from over 20 server performance metrics we are recording historically. When running the client directly from the app server or other RDS servers during these issues, the responsiveness of the client will still be normal.
-
Users are sometimes still on the RDS server after the Kinetics client crashes, which requires manually kicking them off for them to get back in. At least one of the main causes of this seems to be when the UI resource limit is reached - we have already increased the limit with the registry key- but this just changes the error message to “you need to increase the UI limit” to “you are approaching the max UI limit” and the client will start acting unresponsive and close soon after you reach the limit.
-
It seems like once one user starts having issues, many other users will start having the same issue. For example, the white screen issue that is fixed by clearing the client cache, will start happening for many users at once as if the cache is being file locked by another user who initially had an error on the same RDS server, and we clear the entire cache regularly. We have since setup an alternate cache directory and this seems to help, but we still have issues where one user’s error seems to affect others on the same server very often.
-
Users will get the ‘system out of memory’ error even though the server’s memory appears normal, this is supposed to be fixed with the 64 bit version of Kinetics, but we discovered after talking with support that the 64 bit version had even more issues, and we did seem to experience more of these same types of issues while we were testing the 64 bit version. Once one user gets the system out of memory error, often many other users on the same server will start getting this error.
Unfortunately a lot of what we’re seeing is symptoms and we’re not seeing much for errors in event viewer or in the Kinetics trace logs.