Can someone explain this to me?  We have 700 users, so having to do a compile and deploy to fix this issue is insane.  What's the deal with this?  Does anyone know what causes this?

We support multiple disconnected remotes and use roles to manage multiple main details, Every time a Role change is processed we need to purge the vfs settings, then redeploy. For remotes timing is critical so we push remote tasks to Kill ServerHost, purge the db settings, redeploy manually and restart ServerHost. It is painful. I had logged a defect against 7.5.1 asking for a reset on the options page. All it needed to do is kill the users profile settings (simple sql delete) then exit slxclient without rewriting any bad data during the logout. During the next login, they would be reset. Haven't seen it yet

My suggestion to you is to add a clear profile cache button to the login screen. It would push the delete statements directly into the db via sql, then see if it fixes the issue when they login. It probably would not work if their last session has not timed out.