Commit Graph

1905 Commits

Author SHA1 Message Date
md_5
b7ffd83675 Add proper thread safety. Please let me know if this deadlocks your server. 2013-02-10 18:35:09 +11:00
md_5
36f3a7a7f2 If you are a server list owner, and you have not fixed your shit, I kindly request that you do so NOW. I will not waste another 5 hours of my life covering up for you. 4 months is more than long enough for 3 lines of code change. 2013-02-10 16:50:54 +11:00
md_5
cb11c55f81 Add volatile to be sure the JVM wont cache the value 2013-02-10 15:12:35 +11:00
md_5
4b4eea9e4b Fix occasionally showing as 1.3 2013-02-10 15:04:27 +11:00
md_5
27bed04351 Revert "Reduce legacy ping threshold to 1000ms"
This reverts commit d2bb5e9ad6.
2013-02-10 14:52:14 +11:00
md_5
d2bb5e9ad6 Reduce legacy ping threshold to 1000ms 2013-02-10 14:35:13 +11:00
md_5
7a9eda4b61 Fix legacy pinging with some uber ninja coding. 2013-02-10 14:26:59 +11:00
md_5
4542cc4a23 Fix support for windows xp / server 2003 2013-02-10 12:35:52 +11:00
md_5
b3e5c97f6b It came to me in a dream - literally. This should fix all known issues. 2013-02-10 11:54:49 +11:00
md_5
60cb2f461b Tweak to thread pool and safety. 2013-02-10 10:30:58 +11:00
md_5
86bb33ece7 Clearing the queue when they are disconnected might actually fix some issues. 2013-02-10 09:33:31 +11:00
md_5
ff911e1872 Add extra info regarding why a user is disconnected. Should help us troubleshoot these errors - please post the full DC line if reporting a bug. 2013-02-10 08:55:18 +11:00
md_5
0dd4f55fd9 *break when dc'd 2013-02-09 23:00:03 +11:00
md_5
3e21bba4ee Remove debug code now that we know it can be called. 2013-02-09 22:54:14 +11:00
md_5
0dd0ebb15f Run disconnect code on the main thread. 2013-02-09 21:55:33 +11:00
md_5
9bca4e4b70 Add debug to see if statement is ever true 2013-02-09 21:28:34 +11:00
md_5
5e80c9de0b Give a pot shot at bringing network logic a bit closer to CB/Vanilla 2013-02-09 21:05:34 +11:00
md_5
3302eedef5 Small cleanup to netty code, no functional changes. 2013-02-09 20:58:03 +11:00
md_5
a2ffb31ddb Don't scare users with debug mode. 2013-02-09 20:42:08 +11:00
md_5
41579ff315 Add debug logging + more cleanup 2013-02-09 20:27:25 +11:00
md_5
f4572d5786 Nullcheck channel just in case, only null channel on disconnect. 2013-02-09 20:16:23 +11:00
md_5
56d3bd5f39 Use a synced arraylist 2013-02-09 20:10:34 +11:00
md_5
d212863aa7 Implement an uber efficient network engine based on the Java NIO framework Netty. This is basically a complete rewrite of the Minecraft network engine with many distinct advantages. First and foremost, there will no longer be the horrid, and redundant case of 2, or even at times, 3 threads per a connection. Instead low level select/epoll based NIO is used. The number of threads used for network reading and writing will scale automatically to the number of cores for use on your server. In most cases this will be around 8 threads for a 4 core server, much better than the up to 1000 threads that could be in use at one time with the old engine. To facilitate asynchronous packet sending or receiving (currently only chat), a cached thread pool is kept handy. Currently this pool is unbounded, however at this stage we do not need to worry about servers being resource starved due to excessive spam or chat.
== Plugin incompatibilities
As a side effect of this change, plugins which rely on very specific implementation level details within Minecraft are broken. At this point in time, TagAPI and ProtocolLib are affected. If you are a user of ProtocolLib you are advised to update to the latest build, where full support is enabled. If you are a user of TagAPI, support has not yet been added, so you will need to install the updated ProtocolLib so that TagAPI may use its functions.
== Stability
The code within this commit has been very lightly tested in production (300 players for approximately 24 hours), however it is not guaranteed to be free from all bugs. If you experence weird connection behaviour, reporting the bug and steps to reproduce are advised. You are also free to downgrade to the latest recommend build, which is guaranteed to be stable.
== Summary
This commit provides a reduction in threads, which gives the CPU / operating system more time to allocate to the main server threads, as well as various other side benefits such as chat thread pooling and a slight reduction in latency.
This commit is licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported license.
2013-02-09 17:16:27 +11:00
md_5
ed84749d6e Split TPS patch out correctly. 2013-02-08 17:09:51 +11:00
md_5
a1e8ebf5b9 Please people, report to Spigot, not me 2013-02-07 20:23:31 +11:00
md_5
9a4bc05358 Send maps even less often, still causing lag. Appreciate if anyone can PR a config option for this, bit busy atm. 2013-02-06 21:08:58 +11:00
EcoCityCraft
b864492e7a Automatically update upstream versions. Bukkit-1.4.7-R1.0-2-g23f5a05, CraftBukkit-1.4.7-R1.0-5-g6c77179 2013-02-04 18:56:19 -06:00
EcoCityCraft
28fbbb823b Automatically update upstream versions. Bukkit-1.4.7-R1.0-2-g23f5a05, CraftBukkit-1.4.7-R1.0-1-gf6e394d 2013-02-03 04:11:52 -06:00
md_5
72acec7f20 Move tick loop changes to own patch with a few optimizations. 2013-02-03 12:28:39 +11:00
md_5
7b4bba6b3b Actually use the parent pom. 2013-02-03 09:54:16 +11:00
md_5
5853e18f40 Just use the parent pom. 2013-02-03 09:48:23 +11:00
md_5
277b59e4b9 Update POM changes to individual patches. 2013-02-03 09:38:54 +11:00
md_5
841693a0c0 Detect, remove and warn about null tile entities. This prevents odd server crashes, caused by unknown reasons. 2013-02-03 09:21:35 +11:00
md_5
7d858839a7 Send maps, just a lot less often. 2013-02-02 19:41:11 +11:00
md_5
5efea6f794 Only nag if we are finding for the first time.. 2013-02-02 18:11:59 +11:00
md_5
b8994aff61 Abuse plugins which embed metrics in the default location. 2013-02-02 16:41:52 +11:00
md_5
37dc2fc1c4 Plugins break when the intended behaviour is put into place, because for some reason they are all packaging the same classes, an inherently unsafe operation in the case that the classes are loaded at the same time by multiple classloaders - hence we are removing safety from the loaders for now until a correct solution can be put in place. 2013-02-02 16:30:39 +11:00
md_5
87ac58c259 Another attempt at 'fixing' the mob spawn issues present in CraftBukkit. 2013-02-02 09:02:14 +11:00
md_5
cd35b5e5b0 Update to CraftBukkit 1.4.7-R1 2013-01-31 16:56:47 +11:00
EcoCityCraft
b809164b5d Automatically update upstream versions. Bukkit-1.4.7-R1.0-1-gf598dac, CraftBukkit-1.4.7-R1.0-1-gf6e394d 2013-01-30 22:51:16 -06:00
EcoCityCraft
c7057aa887 Automatically update upstream versions. Bukkit-1.4.7-R1.0, CraftBukkit-1.4.7-R1.0 2013-01-30 22:41:17 -06:00
EcoCityCraft
01a0395fa3 Automatically update upstream versions. Bukkit-1.4.7-R0.1-6-g3fc631f, CraftBukkit-1.4.7-R0.1-26-gabccf54 2013-01-30 00:21:30 -06:00
EcoCityCraft
35915ce6cf Automatically update upstream versions. Bukkit-1.4.7-R0.1-6-g3fc631f, CraftBukkit-1.4.7-R0.1-23-gf07dbe2 2013-01-29 07:51:18 -06:00
md_5
a65fedb3f8 Mob spawning issues - 'fix'. See below for ideal reasoning from MikePrimm, however until ideal reasoning we must live with the CraftBukkit / Vanilla behaviour since this causes far too many issues.
IIRC, the main item I was driving towards was a consequence of persistent passive mobs - specifically, the fact that allowing a population limit of N (independent of view distance, which is what vanilla does) when your view distance limits actual loaded chunks to a much smaller area than default (say a view of 4, which would be 9 x 9 chunks loaded per player - and spawnable - versus default, which is 8 radius spawn, or 17 x 17 chunks) tends to result in more mobs per chunk. For persistent mobs, this is bad - since they count for server load and for population just by being loaded, versus being despawned beyond 128 blocks (8 chunk radius - unconditional of view distance, as well) - so they can cause the population limit to be reached more easily, cutting off spawning nearer to players. The goal was to make it so that the mobs-per-loaded-chunk was about the same for all view distances, versus having low view distances cause higher mob concentrations.

    Now, all of this assumes that loaded chunks beyond those around players are modest (since they could contain passive mobs that would count towards the limits) - which they should be, except that recent CBs leak chunks like mad, from what I can see (chunk-gc has become more required than optional), and I think there may be some issues with even hostile mobs "lurking" around - possibly even after their chunks are unloaded. Anything that causes more mobs to be in places players don't see them is going to drive population limit issues, and resulting low spawn behaviors. The trick for us, trying to make big servers as practical as possible, is to shift the math the other way - given low view distances, how to best make sure that folks get reasonable spawn behavior while minimizing the time/resources spent on the server on mobs that don't help that. Realistically, I think we need to analyse the mob demographics better - especially as it relates to lower view distances (our changes have no net impact on view distances above 7) - particularly to understand how the proportion of "useful" mobs is working out (ones close enough to players to be considered contributing to game play). One thought is to manage the population limit based on mobs that are 'tickable' - if they aren't close enough to be ticking, they aren't interesting. This is likely a big issue for view distance 5 folks, since mobs cannot spawn closer than 24 (approx 1 chunk radius, given that middle chunk is 0), and don't tick when any of the chunks within a 2 chunk radius aren't loaded) - so, effectively, they "live" within a zone of 7 x 7 with the middle 3 x 3 removed (so, about 40 chunks) out of a total load zone of 11 x 11 (121) - so about 2/3 of the area containing mobs has idle mobs. Normal view distance would result in mobs ticking as far out as they can spawn (radius 8 versus a load radius of 11), so 100% of the mobs that spawn are ticking when they spawn, and all hostile mobs that are loaded are ticking (since they always despawn beyond 128 blocks / 8 chunks from a player). One interesting thought would be to limit the chunks we spawn mobs in to those where they would be ticking initially (that is, view-distance minus 2 or 3).
2013-01-29 16:57:06 +11:00
EcoCityCraft
165c72a896 Automatically update upstream versions. Bukkit-1.4.7-R0.1-6-g3fc631f, CraftBukkit-1.4.7-R0.1-21-geea7fc2 2013-01-27 23:05:48 -06:00
EcoCityCraft
c686f927a4 Automatically update upstream versions. Bukkit-1.4.7-R0.1-6-g3fc631f, CraftBukkit-1.4.7-R0.1-20-gdeda98a 2013-01-27 17:50:51 -06:00
md_5
2124e746b7 10 years later... 2013-01-28 09:46:19 +11:00
EcoCityCraft
a645cd688d Automatically update upstream versions. Bukkit-1.4.7-R0.1-6-g3fc631f, CraftBukkit-1.4.7-R0.1-19-g4aae815 2013-01-27 11:36:36 -06:00
EcoCityCraft
4b6eabfc58 Automatically update upstream versions. Bukkit-1.4.7-R0.1-6-g3fc631f, CraftBukkit-1.4.7-R0.1-16-g226204e 2013-01-26 13:56:36 -06:00
md_5
3faf5e20e5 Fix tripwires + minecarts 2013-01-26 08:33:09 +11:00