Hello there, I've had an issue with our server for some time now, but not come across any solutions in searches and threads here. Between an hour and an hour and 50 minutes of server runtime all players on our server will notice the server has stopped responding, and get disconnected with an error message "Internal exception: java.net.SocketTimeoutException: Read timed out". There doesn't appear to be any trigger patten. It happens when I'm playing there alone, and I'm never doing the same thing or in the same place when it happens. Players do not have to be present for the issue to happen at all. If the server has been running for say 2 hours and someone tries to join, they immediately get stuck at "downloading terrain". The server still appears to be running normally if I use an external server query tool. It will report the number of player slots, world names, and the players that were connected are still on it. If players attempt to reconnect they will be stuck on the "downloading terrain" prompt. Using a 3rd party mod to show my co-ordinates at the top left (replaces version string), I see that I'm falling until it notices the server isn't sending data and gets "read timed out". I'm finding this hard to troubleshoot as the logs do not show any clue as to why this happens. It just looks normal. As it's a rented server I can't see a console window. Solution It turned out that iConomy 5.0/5.0.1 was the direct cause of this issue. Details Server information Hardware: Four Intel quad-core Xeon's with many GB of RAM (1.5 GB allocated to our Minecraft server process), stored on an SSD. (shared server with insane hardware) OS: FreeBSD 8.2 Release P2 amd64. Java: OpenJDK based on JDK 1.6, probably also amd64. Wrappers: None. Server build: CraftBukkit build 1060 on Minecraft 1.7.3, though this issue happened with builds 1000, 953, and any other that were released since Minecraft 1.6 Launch command: Unsure, as it's rented and started/stopped/monitored from a web page (not on the same box as the Minecraft server). Plugins: Most of these were present long before we had this issue. AdminCmd 5.7.1 AppleTree 0.5 AutoSave 2.0.1 (installed after the issue, just to ensure we don't lose any data) ChestShop 3.01 iChat 2.3.1 (replaced for mChat due to new permissions base) iConomy 5.01 iConomyChestShop 2.75 (replaced for ChestShop 3.01) LWC 3.31 mChat 1.7.10_5 MCDocs 10 nChat 0.5 was removed as we thought this may be causing the problem (player chat was no longer getting formatted around the same time the issue started), replaced it with iChat NoCheat 1.11 Permissions 3.0a (replaced for PermissionsBukkit and SuperpermsBridge) PermissionsBukkit 1.1 PermissionsPlus 2.1.1 (replaced for PermissionsBukkit and SuperpermsBridge) Remote Bukkit 0.4 SuperpermsBridge 1.2 Stackable 0.5.1 StarGate 0.5.5 Towny 0.74 Warpz0r 1.3.1 WorldEdit 4.6 WorldGuard 5.2.2 Error log: Nothing at all shows any indication around the time the mass time out occurs. All that happens is a few minutes after it happens and the players are shown the "read timed out", a few generic connection timeout messages appear for players. That would be normal as the players are no longer accepting data from the server after they've left. Resolution attempts: First it was suggested our world could be too large or have "errors", so I used MCEdit to trim out chunks nobody used (like when someone just wanders off for miles for no reason). That didn't help, so I changed to use a fresh world -- no change. It was also suggested that the originally allocated 600MB of RAM was not sufficient for 8 players with the plug-ins we had, so I used the MemMonitor plug-in to verify that. With 1 person connected and doing stuff the average usage was 280 MB out of 587 MB allocated. Even with 5 of us there it didn't reach 500 MB. We were upgraded to 1.5 GB to make sure it wasn't a memory usage problem. Later I noticed that iConomy was posting "[INFO] [iConomy] Could not create connection: java.sql.SQLException: Login timeout" every minute or so in the server log even though we use H2SQL. This was confirmed to be a bug in iConomy 5.0, and updated to 5.01 fixed this. As we have FTP access to the Minecraft server folder I downloaded it all and ran it locally. I joined it and left myself connected for a few hours while I went to the pub for lunch. Came back 3 hours later, it was running just fine with me still standing there. I am running Sun's JDK 1.7 x64 on my machines though. Oddly with just me connected it used 160 MB less RAM than alone on our rented server, only averaging 120 MB usage. Finally, as the server host has many servers we were moved to a different machine -- same issue. Resolution suggestions: I have not tried running vanilla yet, mainly because switching to it erases our plug-in configuration and data files -- and there is a lot of it. The same with removing plug-ins individually. I will try this if no other suggestion is available, but I'm hoping that one of the plug-ins we use may be known to cause problems. Thank you for your time reading all of that. I will be grateful for any suggestions, help, or information you may have. I really can't see where the issue is.