Flink学习 - 4. Flink 配置文件参数 && Flink任务提交参数

配置文件说明

摘自： Flink从0到1学习 – Flink配置文件详解

官网介绍： https://ci.apache.org/projects/flink/flink-docs-stable/ops/config.html

安装目录下主要有 flink-conf.yaml 配置、日志的配置文件、zk 配置、Flink SQL Client 配置。

flink-conf.yaml

基础配置

# jobManager 的IP地址
jobmanager.rpc.address: localhost

# JobManager 的端口号
jobmanager.rpc.port: 6123

# JobManager JVM heap 内存大小
jobmanager.heap.size: 1024m

# TaskManager JVM heap 内存大小
taskmanager.heap.size: 1024m

# 每个 TaskManager 提供的任务 slots 数量大小

taskmanager.numberOfTaskSlots: 1

# 程序默认并行计算的个数
parallelism.default: 1

# 文件系统来源
# fs.default-scheme

高可用性配置

# 可以选择 'NONE' 或者 'zookeeper'.
# high-availability: zookeeper

# 文件系统路径，让 Flink 在高可用性设置中持久保存元数据
# high-availability.storageDir: hdfs:///flink/ha/

# zookeeper 集群中仲裁者的机器 ip 和 port 端口号
# high-availability.zookeeper.quorum: localhost:2181

# 默认是 open，如果 zookeeper security 启用了该值会更改成 creator
# high-availability.zookeeper.client.acl: open

容错和检查点 配置

# 用于存储和检查点状态
# state.backend: filesystem

# 存储检查点的数据文件和元数据的默认目录
# state.checkpoints.dir: hdfs://namenode-host:port/flink-checkpoints

# savepoints 的默认目标目录(可选)
# state.savepoints.dir: hdfs://namenode-host:port/flink-checkpoints

# 用于启用/禁用增量 checkpoints 的标志
# state.backend.incremental: false

web 前端配置

# 基于 Web 的运行时监视器侦听的地址.
#jobmanager.web.address: 0.0.0.0

#  Web 的运行时监视器端口
rest.port: 8081

# 是否从基于 Web 的 jobmanager 启用作业提交
# jobmanager.web.submit.enable: false

高级配置

# io.tmp.dirs: /tmp

# 是否应在 TaskManager 启动时预先分配 TaskManager 管理的内存
# taskmanager.memory.preallocate: false

# 类加载解析顺序，是先检查用户代码 jar（“child-first”）还是应用程序类路径（“parent-first”）。 默认设置指示首先从用户代码 jar 加载类
# classloader.resolve-order: child-first


# 用于网络缓冲区的 JVM 内存的分数。 这决定了 TaskManager 可以同时拥有多少流数据交换通道以及通道缓冲的程度。 如果作业被拒绝或者您收到系统没有足够缓冲区的警告，请增加此值或下面的最小/最大值。 另请注意，“taskmanager.network.memory.min”和“taskmanager.network.memory.max”可能会覆盖此分数

# taskmanager.network.memory.fraction: 0.1
# taskmanager.network.memory.min: 67108864
# taskmanager.network.memory.max: 1073741824

Flink 集群安全配置

# 指示是否从 Kerberos ticket 缓存中读取
# security.kerberos.login.use-ticket-cache: true

# 包含用户凭据的 Kerberos 密钥表文件的绝对路径
# security.kerberos.login.keytab: /path/to/kerberos/keytab

# 与 keytab 关联的 Kerberos 主体名称
# security.kerberos.login.principal: flink-user

# 以逗号分隔的登录上下文列表，用于提供 Kerberos 凭据（例如，`Client，KafkaClient`使用凭证进行 ZooKeeper 身份验证和 Kafka 身份验证）
# security.kerberos.login.contexts: Client,KafkaClient

Zookeeper 安全配置

# 覆盖以下配置以提供自定义 ZK 服务名称
# zookeeper.sasl.service-name: zookeeper

# 该配置必须匹配 "security.kerberos.login.contexts" 中的列表（含有一个）
# zookeeper.sasl.login-context-name: Client

HistoryServer

# 你可以通过 bin/historyserver.sh (start|stop) 命令启动和关闭 HistoryServer

# 将已完成的作业上传到的目录
# jobmanager.archive.fs.dir: hdfs:///completed-jobs/

# 基于 Web 的 HistoryServer 的地址
# historyserver.web.address: 0.0.0.0

# 基于 Web 的 HistoryServer 的端口号
# historyserver.web.port: 8082

# 以逗号分隔的目录列表，用于监视已完成的作业
# historyserver.archive.fs.dir: hdfs:///completed-jobs/

# 刷新受监控目录的时间间隔（以毫秒为单位）
# historyserver.archive.fs.refresh-interval: 10000

masters

host：port

localhost:8081

slaves

每个worker节点的IP/Hostname

localhost

zoo.cfg

# 每个 tick 的毫秒数
tickTime=2000

# 初始同步阶段可以采用的 tick 数
initLimit=10

# 在发送请求和获取确认之间可以传递的 tick 数
syncLimit=5

# 存储快照的目录
# dataDir=/tmp/zookeeper

# 客户端将连接的端口
clientPort=2181

# ZooKeeper quorum peers
server.1=localhost:2888:3888
# server.2=host:peer-port:leader-port

日志配置

Flink 在不同平台下运行的日志文件

log4j-cli.properties
log4j-console.properties
log4j-yarn-session.properties
log4j.properties
logback-console.xml
logback-yarn.xml
logback.xml

sql-client-defaults.yaml

execution:
  # 'batch' or 'streaming' execution
  type: streaming
  # allow 'event-time' or only 'processing-time' in sources
  time-characteristic: event-time
  # interval in ms for emitting periodic watermarks
  periodic-watermarks-interval: 200
  # 'changelog' or 'table' presentation of results
  result-mode: changelog
  # parallelism of the program
  parallelism: 1
  # maximum parallelism
  max-parallelism: 128
  # minimum idle state retention in ms
  min-idle-state-retention: 0
  # maximum idle state retention in ms
  max-idle-state-retention: 0
  
deployment:
  # general cluster communication timeout in ms
  response-timeout: 5000
  # (optional) address from cluster to gateway
  gateway-address: ""
  # (optional) port from cluster to gateway
  gateway-port: 0

参数说明

执行 flink run 后参数：

参数说明
Action "run" compiles and runs a program.
Syntax: run [OPTIONS]  

"run" action options:
 -c,--class                          Class with the program entry
                                                point ("main" method or
                                                "getPlan()" method. Only
                                                needed if the JAR file does
                                                not specify the class in its
                                                manifest.
 -C,--classpath                            Adds a URL to each user code
                                                classloader  on all nodes in
                                                the cluster. The paths must
                                                specify a protocol (e.g.
                                                file://) and be accessible
                                                on all nodes (e.g. by means
                                                of a NFS share). You can use
                                                this option multiple times
                                                for specifying more than one
                                                URL. The protocol must be
                                                supported by the {@link
                                                java.net.URLClassLoader}.
 -d,--detached                                  If present, runs the job in
                                                detached mode
 -m,--jobmanager                     Address of the JobManager
                                                (master) to which to
                                                connect. Use this flag to
                                                connect to a different
                                                JobManager than the one
                                                specified in the
                                                configuration.
 -p,--parallelism                  The parallelism with which
                                                to run the program. Optional
                                                flag to override the default
                                                value specified in the
                                                configuration.
 -q,--sysoutLogging                             If present, suppress logging
                                                output to standard out.
 -s,--fromSavepoint              Path to a savepoint to reset
                                                the job back to (for example
                                                file:///flink/savepoint-1537
                                                ).
 -z,--zookeeperNamespace    Namespace to create the
                                                Zookeeper sub-paths for high
                                                availability mode
Options for yarn-cluster mode:
 -yD                             Dynamic properties
 -yd,--yarndetached                   Start detached
 -yid,--yarnapplicationId        Attach to running YARN session
 -yj,--yarnjar                   Path to Flink jar file
 -yjm,--yarnjobManagerMemory     Memory for JobManager Container [in
                                      MB]
 -yn,--yarncontainer             Number of YARN container to allocate
                                      (=Number of Task Managers)
 -ynm,--yarnname                 Set a custom name for the application
                                      on YARN
 -yq,--yarnquery                      Display available YARN resources
                                      (memory, cores)
 -yqu,--yarnqueue                Specify YARN queue.
 -ys,--yarnslots                 Number of slots per TaskManager
 -yst,--yarnstreaming                 Start Flink in streaming mode
 -yt,--yarnship                  Ship files in the specified directory
                                      (t for transfer)
 -ytm,--yarntaskManagerMemory    Memory per TaskManager Container [in
                                      MB]
 -yz,--yarnzookeeperNamespace    Namespace to create the Zookeeper
                                      sub-paths for high availability mode

-n 10 一共启动10个TaskManager节点

-jm 1024 JobManager的内存大小为1024M
 
-tm 2048 TaskManager的内存大小为2048M
-d 使用detached模式进行部署（部署完成后本地命令行可以退出）
-qu default 部署到YARN的default队列中