Detect when the server has been hung for a long duration, and start printing
thread dumps at an interval until the point of crash.
This will help diagnose what was going on in that time before the crash.
The watchdog thread calls the server restart function asynchronously. Prior to
this change, it attempted to do several non-safe operations from the watchdog
thread, rather than the main. Specifically, because of a separate upstream change,
it causes player entities to be ticked asynchronously, among other things.
This is dangerous.
This patch moves the old handling into a synchronous variant, for calls from the
restart command, and adds separate handling for async calls, such as those from
the watchdog thread.
When calling from the watchdog thread, we cannot assume the main thread is in a
tickable state; it may be completely deadlocked. In order to handle this, we mark
the server as stopping, in order to account for situations where the server should
complete a tick reasonbly soon, i.e. 99% of cases.
Should the server not enter a state where it is stopping within 10 seconds, We
will assume that the server has in fact deadlocked and will proceed to force
kill the server.
This modification does not force restart the server should we actually enter a
deadlocked state where the server is stopping, whereas this will in most cases
exit within a reasonable amount of time, to put a fixed limit on a process that
will have plugins and worlds saving to the disk has a high potential to result
in corruption/dataloss.
The process of determining an entity's exposure from explosions can be
expensive when there are hundreds or more entities in range.
This patch adds a per-tick cache that is used for storing and retrieving
an entity's exposure during an explosion.
Improves how the catchup buffer is handled, allowing it to roll both ways
increasing the effeciency of the thread sleep so it only will sleep once.
Also increases the buffer of the catchup to ensure server stays at 20 TPS unless extreme conditions
Previous implementation did not calculate TPS correctly.
Switch to a realistic rolling average and factor in std deviation as an extra reporting variable
Rewrite console improvements (console colors, tab completion,
persistent input line, ...) using JLine 3.x and TerminalConsoleAppender.
Also uses the new ANSIComponentSerializer to serialize components when
logging them via the ComponentLogger, or when sending messages to the
console, for hex color support.
New features:
- Support console colors for Vanilla commands
- Add console colors for warnings and errors
- Server can now be turned off safely using CTRL + C. JLine catches
the signal and the implementation shuts down the server cleanly.
- Support console colors and persistent input line when running in
IntelliJ IDEA
Other changes:
- Server starts 1-2 seconds faster thanks to optimizations in Log4j
configuration
Co-Authored-By: Emilia Kond <emilia@rymiel.space>
== AT ==
public net.minecraft.network.chat.HoverEvent$ItemStackInfo item
public net.minecraft.network.chat.HoverEvent$ItemStackInfo count
public net.minecraft.network.chat.HoverEvent$ItemStackInfo components
public net.minecraft.network.chat.contents.TranslatableContents filterAllowedArguments(Ljava/lang/Object;)Lcom/mojang/serialization/DataResult;
Co-authored-by: zml <zml@stellardrift.ca>
Co-authored-by: Jake Potrebic <jake.m.potrebic@gmail.com>
== AT ==
public net.minecraft.server.level.ServerChunkCache mainThread
public net.minecraft.server.level.ServerLevel chunkSource
public org.bukkit.craftbukkit.inventory.CraftItemStack handle
public net.minecraft.server.level.ChunkMap getVisibleChunkIfPresent(J)Lnet/minecraft/server/level/ChunkHolder;
public net.minecraft.server.level.ServerChunkCache mainThreadProcessor
public net.minecraft.server.level.ServerChunkCache$MainThreadExecutor
public net.minecraft.world.level.chunk.LevelChunkSection states
== AT ==
public org.spigotmc.SpigotWorldConfig getBoolean(Ljava/lang/String;Z)Z
public org.spigotmc.SpigotWorldConfig getDouble(Ljava/lang/String;)D
public org.spigotmc.SpigotWorldConfig getDouble(Ljava/lang/String;D)D
public org.spigotmc.SpigotWorldConfig getInt(Ljava/lang/String;)I
public org.spigotmc.SpigotWorldConfig getInt(Ljava/lang/String;I)I
public org.spigotmc.SpigotWorldConfig getList(Ljava/lang/String;Ljava/lang/Object;)Ljava/util/List;
public org.spigotmc.SpigotWorldConfig getString(Ljava/lang/String;Ljava/lang/String;)Ljava/lang/String;
public net.minecraft.server.dedicated.DedicatedServerProperties reload(Lnet/minecraft/core/RegistryAccess;Ljava/util/Properties;Ljoptsimple/OptionSet;)Lnet/minecraft/server/dedicated/DedicatedServerProperties;
public net.minecraft.world.level.NaturalSpawner SPAWNING_CATEGORIES
This brings back the option that the Spigot version of netty saw. By default Netty will try and use cores*2 threads, however if running multiple servers on the same machine, this can be too many threads. Additionally some people have 16 core servers. If 32 Netty threads are allowed in this setup, then the lock contention, and thus blocking between threads becomes much greater, leading to decreased performance.
By: md_5 <git@md-5.net>
