vmarkis

I'm trying to reconnect the user and I'm handling all possible errors. As you can see in the logs, the first module I wrote in the stack trace is `SteamAccountClient.login()`, which calls `client.logOn() (a method from steam-user)` This image shows the code for `SteamAccountClient.login()`. It's wrapping `client.logOn()` in a try-catch block, so no error should've be thrown. I noticed an internal use of the `logOn()` method within a timeout, but without any try-catch block. This could be why my app is crashing, even though it's shown in the stack trace.

Let's say I have 1000 active users on my server. I deploy some changes and need to apply into the server. All these 1000 active instances are on the process memory. If I restart the server the instances are gone. I've seen McKay talking about splitting the app into multiple process so it becomes less prone to changes. Say these uses are all farming hours in games, how should I reconnect these 1000 users back again? Without causing any problems like 1000 people connecting from the same 'origin' at the same time. Should I split the active accounts into small groups and reconnecting them one group per time, with a delay between them?