无法将Log4J中的事件放入Flume

Question

我正在尝试使用Log4J Flume appender通过Flume将事件从Log4J 1x获取到HDFS。创建了两个附加器FILE和水槽。它适用于文件附加器，但使用水槽附加器，程序只是挂在Eclipse中。Flume工作正常，我能够使用avro客户端向avro源发送消息并在HDFS中查看消息。但是，它没有与Log4J 1x集成。

我没有看到任何异常，除了下面在log.out中。

Batch size string = null
Using Netty bootstrap options: {tcpNoDelay=true, connectTimeoutMillis=20000}
Connecting to localhost/127.0.0.1:41414
[id: 0x52a00770] OPEN

从水槽控制台

2013-10-23 14:32:32,145 (pool-5-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.handleUpstream(NettyServer.java:171)] [id: 0x577cf6e4, /127.0.0.1:46037 => /127.0.0.1:41414] OPEN
2013-10-23 14:32:32,148 (pool-6-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.handleUpstream(NettyServer.java:171)] [id: 0x577cf6e4, /127.0.0.1:46037 => /127.0.0.1:41414] BOUND: /127.0.0.1:41414
2013-10-23 14:32:32,148 (pool-6-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.handleUpstream(NettyServer.java:171)] [id: 0x577cf6e4, /127.0.0.1:46037 => /127.0.0.1:41414] CONNECTED: /127.0.0.1:46037
2013-10-23 14:32:43,086 (pool-6-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.handleUpstream(NettyServer.java:171)] [id: 0x577cf6e4, /127.0.0.1:46037 :> /127.0.0.1:41414] DISCONNECTED
2013-10-23 14:32:43,096 (pool-6-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.handleUpstream(NettyServer.java:171)] [id: 0x577cf6e4, /127.0.0.1:46037 :> /127.0.0.1:41414] UNBOUND
2013-10-23 14:32:43,096 (pool-6-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.handleUpstream(NettyServer.java:171)] [id: 0x577cf6e4, /127.0.0.1:46037 :> /127.0.0.1:41414] CLOSED
2013-10-23 14:32:43,097 (pool-6-thread-1) [INFO - org.apache.avro.ipc.NettyServer$NettyServerAvroHandler.channelClosed(NettyServer.java:209)] Connection to /127.0.0.1:46037 disconnected.

如果有帮助的话，我确实在调试模式下运行了程序，当它挂起时，我会挂起并进行堆栈跟踪。试图查看代码，但不确定为什么程序会挂起水槽附加器。

Daemon Thread [Avro NettyTransceiver I/O Worker 1] (Suspended)  
Logger(Category).callAppenders(LoggingEvent) line: 205  
Logger(Category).forcedLog(String, Priority, Object, Throwable) line: 391  
Logger(Category).log(String, Priority, Object, Throwable) line: 856  
Log4jLoggerAdapter.debug(String) line: 209  
NettyTransceiver$NettyClientAvroHandler.handleUpstream(ChannelHandlerContext, ChannelEvent) line: 491  
DefaultChannelPipeline.sendUpstream(DefaultChannelPipeline$DefaultChannelHandlerContext, ChannelEvent) line: 564  
DefaultChannelPipeline$DefaultChannelHandlerContext.sendUpstream(ChannelEvent) line: 792  
NettyTransportCodec$NettyFrameDecoder(SimpleChannelUpstreamHandler).channelBound(ChannelHandlerContext, ChannelStateEvent) line: 166  
NettyTransportCodec$NettyFrameDecoder(SimpleChannelUpstreamHandler).handleUpstream(ChannelHandlerContext, ChannelEvent) line: 98  
DefaultChannelPipeline.sendUpstream(DefaultChannelPipeline$DefaultChannelHandlerContext, ChannelEvent) line: 564  
DefaultChannelPipeline.sendUpstream(ChannelEvent) line: 559  
Channels.fireChannelBound(Channel, SocketAddress) line: 199  
NioWorker$RegisterTask.run() line: 191  
NioWorker(AbstractNioWorker).processRegisterTaskQueue() line: 329  
NioWorker(AbstractNioWorker).run() line: 235  
NioWorker.run() line: 38  
DeadLockProofWorker$1.run() line: 42  
ThreadPoolExecutor.runWorker(ThreadPoolExecutor$Worker) line: 1145  
ThreadPoolExecutor$Worker.run() line: 615  
Thread.run() line: 744

这是Java程序

import java.io.IOException;
import java.sql.SQLException;
import org.apache.log4j.Logger;
public class log4jExample {
    static Logger log = Logger.getRootLogger();
    public static void main(String[] args) throws IOException, SQLException {
       log.debug("Hello this is an debug message");
    }
}

下面是log4j.properties

# Define the root logger with appender file
log = /home/vm4learning/WorkSpace/BigData/Log4J-Example/log
log4j.rootLogger = DEBUG, FILE, flume

# Define the file appender
log4j.appender.FILE=org.apache.log4j.FileAppender
log4j.appender.FILE.File=${log}/log.out
log4j.appender.FILE.layout=org.apache.log4j.PatternLayout
log4j.appender.FILE.layout.conversionPattern=%m%n

# Define the flume appender
log4j.appender.flume = org.apache.flume.clients.log4jappender.Log4jAppender
log4j.appender.flume.Hostname = localhost
log4j.appender.flume.Port = 41414
log4j.appender.flume.UnsafeMode = false
log4j.appender.flume.layout=org.apache.log4j.PatternLayout
log4j.appender.flume.layout.ConversionPattern=%m%n

以下是Eclipse中的依赖项

flume-ng-log4jappender-1.4.0.jar
log4j-1.2.17.jar
flume-ng-sdk-1.4.0.jar
avro-1.7.3.jar
netty-3.4.0.Final.jar
avro-ipc-1.7.3.jar
slf4j-api-1.6.1.jar
slf4j-log4j12-1.6.1.jar

以下是flume.conf的内容

# Tell agent1 which ones we want to activate.
agent1.channels = ch1
agent1.sources = avro-source1
agent1.sinks = hdfs-sink1

# Define a memory channel called ch1 on agent1
agent1.channels.ch1.type = memory

# Define an Avro source called avro-source1 on agent1 and tell it
# to bind to 0.0.0.0:41414. Connect it to channel ch1.
agent1.sources.avro-source1.type = avro
agent1.sources.avro-source1.bind = 0.0.0.0
agent1.sources.avro-source1.port = 41414

# Define a logger sink that simply logs all events it receives
# and connect it to the other end of the same channel.
agent1.sinks.hdfs-sink1.type = hdfs
agent1.sinks.hdfs-sink1.hdfs.path = hdfs://localhost:9000/flume/events/

agent1.sinks.hdfs-sink1.channel = ch1
agent1.sources.avro-source1.channels = ch1

如何解决这个问题？

Answer 1

我曾经在log4j中使用Flume附加器时遇到过类似的问题。每当我尝试实例化 Logger 对象时，程序都会挂起。我记得问题是我在类路径中没有所有必需的库，一旦我添加了它们，它就可以正常工作。

我建议你先从迈克·珀西的简单例子开始。虽然pom.xml那里创建了一个包含所有依赖项的JAR，但编辑它以将依赖的jar文件复制到另一个目录会给我这个列表：

avro-1.7.4.jar
avro-ipc-1.7.4.jar
commons-codec-1.3.jar
commons-collections-3.2.1.jar
commons-compress-1.4.1.jar
commons-lang-2.5.jar
commons-logging-1.1.1.jar
flume-ng-log4jappender-1.4.0-cdh4.5.0.jar
flume-ng-sdk-1.4.0-cdh4.5.0.jar
hamcrest-core-1.1.jar
httpclient-4.0.1.jar
httpcore-4.0.1.jar
jackson-core-asl-1.8.8.jar
jackson-mapper-asl-1.8.8.jar
jetty-6.1.26.jar
jetty-util-6.1.26.jar
junit-4.10.jar
libthrift-0.7.0.jar
log4j-1.2.16.jar
netty-3.5.0.Final.jar
paranamer-2.3.jar
slf4j-api-1.7.2.jar
slf4j-jdk14-1.7.2.jar
snappy-java-1.0.4.1.jar
velocity-1.7.jar
xz-1.0.jar

其中一些库（如jUnit）可能并不是真正必要的，但我建议首先使用所有这些库，看看您是否可以让您的示例正常工作，然后尝试确定所需的最小集。

Answer 2

我的猜测是，您正在尝试通过Flume记录Flume的事件。我已经在其他附加器中看到过这个问题，但在 Log4j1 上没有。

我会考虑修改log4j.properties以排除Flume，Netty和Avro事件，看看是否可以修复它。

Answer 3

我也遇到过类似的问题，解决方法是：

• 将 log4j.properties 根记录器从调试级别更改为信息级别。

但是我不知道水槽里发生了什么。我正在尝试调试它。如果有人知道，告诉我谢谢~~