kafka 打开的文件太多

Question

你有没有遇到过关于 kafka 的类似问题？我收到此错误：Too many open files。我不知道为什么。以下是一些日志：

[2018-08-27 10:07:26,268] ERROR Error while deleting the clean shutdown file in dir /home/weihu/kafka/kafka/logs (kafka.server.LogD)
java.nio.file.FileSystemException: /home/weihu/kafka/kafka/logs/BC_20180821_1_LOCATION-87/leader-epoch-checkpoint: Too many open fis
        at sun.nio.fs.UnixException.translateToIOException(UnixException.java:91)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107)
        at sun.nio.fs.UnixFileSystemProvider.newByteChannel(UnixFileSystemProvider.java:214)
        at java.nio.file.Files.newByteChannel(Files.java:361)
        at java.nio.file.Files.createFile(Files.java:632)
        at kafka.server.checkpoints.CheckpointFile.<init>(CheckpointFile.scala:45)
        at kafka.server.checkpoints.LeaderEpochCheckpointFile.<init>(LeaderEpochCheckpointFile.scala:62)
        at kafka.log.Log.initializeLeaderEpochCache(Log.scala:278)
        at kafka.log.Log.<init>(Log.scala:211)
        at kafka.log.Log$.apply(Log.scala:1748)
        at kafka.log.LogManager.loadLog(LogManager.scala:265)
        at kafka.log.LogManager.$anonfun$loadLogs$12(LogManager.scala:335)
        at kafka.utils.CoreUtils$$anon$1.run(CoreUtils.scala:62)
        at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
        at java.util.concurrent.FutureTask.run(FutureTask.java:266)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)
[2018-08-27 10:07:26,268] ERROR Error while deleting the clean shutdown file in dir /home/weihu/kafka/kafka/logs (kafka.server.LogD)
java.nio.file.FileSystemException: /home/weihu/kafka/kafka/logs/BC_20180822_PARSE-136/leader-epoch-checkpoint: Too many open files
        at sun.nio.fs.UnixException.translateToIOException(UnixException.java:91)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107)
        at sun.nio.fs.UnixFileSystemProvider.newByteChannel(UnixFileSystemProvider.java:214)
        at java.nio.file.Files.newByteChannel(Files.java:361)
        at java.nio.file.Files.createFile(Files.java:632)
        at kafka.server.checkpoints.CheckpointFile.<init>(CheckpointFile.scala:45)
        at kafka.server.checkpoints.LeaderEpochCheckpointFile.<init>(LeaderEpochCheckpointFile.scala:62)
        at kafka.log.Log.initializeLeaderEpochCache(Log.scala:278)
        at kafka.log.Log.<init>(Log.scala:211)
        at kafka.log.Log$.apply(Log.scala:1748)
        at kafka.log.LogManager.loadLog(LogManager.scala:265)
        at kafka.log.LogManager.$anonfun$loadLogs$12(LogManager.scala:335)
        at kafka.utils.CoreUtils$$anon$1.run(CoreUtils.scala:62)
        at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
        at java.util.concurrent.FutureTask.run(FutureTask.java:266)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)
[2018-08-27 10:07:26,269] ERROR Error while deleting the clean shutdown file in dir /home/weihu/kafka/kafka/logs (kafka.server.LogD)
java.nio.file.FileSystemException: /home/weihu/kafka/kafka/logs/BC_20180813_1_STATISTICS-402/leader-epoch-checkpoint: Too many opens
        at sun.nio.fs.UnixException.translateToIOException(UnixException.java:91)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102)
        at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107)
        at sun.nio.fs.UnixFileSystemProvider.newByteChannel(UnixFileSystemProvider.java:214)
        at java.nio.file.Files.newByteChannel(Files.java:361)
        at java.nio.file.Files.createFile(Files.java:632)
        at kafka.server.checkpoints.CheckpointFile.<init>(CheckpointFile.scala:45)
        at kafka.server.checkpoints.LeaderEpochCheckpointFile.<init>(LeaderEpochCheckpointFile.scala:62)
        at kafka.log.Log.initializeLeaderEpochCache(Log.scala:278)
        at kafka.log.Log.<init>(Log.scala:211)
        at kafka.log.Log$.apply(Log.scala:1748)
        at kafka.log.LogManager.loadLog(LogManager.scala:265)
        at kafka.log.LogManager.$anonfun$loadLogs$12(LogManager.scala:335)
        at kafka.utils.CoreUtils$$anon$1.run(CoreUtils.scala:62)
        at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
        at java.util.concurrent.FutureTask.run(FutureTask.java:266)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)

Answer 1

在 Kafka 中，每个主题都（可选地）分成许多分区。对于每个分区，一些文件由代理维护（用于索引和实际数据）。

kafka-topics --zookeeper localhost:2181 --describe --topic topic_name

将为您提供主题 topic_name 的分区数。每个主题num.partitions 的默认分区数在/etc/kafka/server.properties 下定义

如果代理托管许多分区并且特定分区有许多日志段文件，则打开文件的总数可能非常大。

运行可以看到当前文件描述符的限制

ulimit -n

您也可以使用lsof查看打开的文件数：

lsof | wc -l

要解决此问题，您需要更改打开文件描述符的限制：

ulimit -n <noOfFiles>

或以某种方式减少打开文件的数量（例如，减少每个主题的分区数量）。

Answer 2

在使用 Systemd 的 Linux 发行版（如 RHEL 和 CentOS）上，您需要在 Systemd 服务文件的第二个块中添加配置行。仅更改 /etc/security/limits.conf 是不够的。

[Unit]
Requires=network.target remote-fs.target
After=network.target remote-fs.target

[Service]
Type=simple
User=kafka

LimitAS=infinity
LimitRSS=infinity
LimitCORE=infinity
LimitNOFILE=65536

ExecStart=/home/kafka/kafka/bin/zookeeper-server-start.sh /home/kafka/kafka/config/zookeeper.properties
ExecStop=/home/kafka/kafka/bin/zookeeper-server-stop.sh
Restart=on-abnormal

[Install]
WantedBy=multi-user.target