参考:
https://ververica.cn/developers-resources/
#flink参数
https://blog.csdn.net/qq_35440040/article/details/84992796
spark使用批处理模拟流计算
flink使用流框架模拟批计算
https://ci.apache.org/projects/flink/flink-docs-release-1.8/
https://flink.apache.org/downloads.html#
下载包：
https://flink.apache.org/downloads.html
tar -xzvf flink-1.8.0-bin-scala_2.11.tgz -C /opt/module/
vim /etc/profile
export FLINK_HOME=/opt/module/flink-1.8.0
export PATH=$PATH:$FLINK_HOME/bin
cd /opt/module/flink-1.8.0/conf
mv flink-conf.yaml flink-conf.yaml.bak
vim flink-conf.yaml
jobmanager.rpc.address: Fengfeng-dr-algo1
jobmanager.rpc.port: 6123
jobmanager.heap.size: 1024m
taskmanager.heap.size: 1024m
taskmanager.numberOfTaskSlots: 2
parallelism.default: 2
fs.default-scheme: hdfs://Fengfeng-dr-algo1:9820
#这个是在core-site.xml里配的hdfs集群地址,yarn集群模式主要配这个
vim masters
Fengfeng-dr-algo1
vim slaves
Fengfeng-dr-algo2
Fengfeng-dr-algo3
Fengfeng-dr-algo4
#配置完成后将文件同步到其他节点
scp /etc/profile Fengfeng-dr-algo2:/etc/profile
scp /etc/profile Fengfeng-dr-algo3:/etc/profile
scp /etc/profile Fengfeng-dr-algo4:/etc/profile
scp -r /opt/module/flink-1.8.0/ Fengfeng-dr-algo2:/opt/module
scp -r /opt/module/flink-1.8.0/ Fengfeng-dr-algo3:/opt/module
scp -r /opt/module/flink-1.8.0/ Fengfeng-dr-algo4:/opt/module
启动集群start-cluster.sh
检查TaskManagerRunner服务起来没有：
[root@Fengfeng-dr-algo1 conf]# ansible all -m shell -a 'jps'
Fengfeng-dr-algo3 | SUCCESS | rc=0 >>
20978 DataNode
22386 TaskManagerRunner
22490 Jps
21295 NodeManager
Fengfeng-dr-algo4 | SUCCESS | rc=0 >>
24625 NodeManager
26193 TaskManagerRunner
24180 DataNode
24292 SecondaryNameNode
26297 Jps
Fengfeng-dr-algo2 | SUCCESS | rc=0 >>
26753 Jps
24867 ResourceManager
24356 DataNode
25480 NodeManager
26650 TaskManagerRunner
Fengfeng-dr-algo1 | SUCCESS | rc=0 >>
27216 Jps
24641 NameNode
24789 DataNode
27048 StandaloneSessionClusterEntrypoint
25500 NodeManager
查看WebUI，端口为8081
#运行flink测试,1.txt在hdfs上.
1/ 以standalone模式
flink run /opt/module/flink-1.8.0/examples/batch/WordCount.jar -c wordcount --input /1.txt
2/ 以yarn-cluster模式，需要停掉集群模式stop-cluster.sh
flink run -m yarn-cluster /opt/module/flink-1.8.0/examples/batch/WordCount.jar -c wordcount --input /1.txt
yarn-cluster跑得作业情况可在yarn的web8080端口看
附: flink yarn-cluster跑wordcount结果
[root@fengfeng-dr-algo1 hadoop]# flink run -m yarn-cluster /opt/module/flink-1.8.0/examples/batch/WordCount.jar -c wordcount --input /1.txt
2019-08-15 03:52:50,622 INFO org.apache.hadoop.yarn.client.RMProxy - Connecting to ResourceManager at oride-dr-algo2/172.28.20.168:8032
2019-08-15 03:52:50,755 INFO org.apache.flink.yarn.cli.FlinkYarnSessionCli - No path for the flink jar passed. Using the location of class org.apache.flink.yarn.YarnClusterDescriptor to locate the jar
2019-08-15 03:52:50,755 INFO org.apache.flink.yarn.cli.FlinkYarnSessionCli - No path for the flink jar passed. Using the location of class org.apache.flink.yarn.YarnClusterDescriptor to locate the jar
2019-08-15 03:52:50,922 WARN org.apache.flink.yarn.AbstractYarnClusterDescriptor - Neither the HADOOP_CONF_DIR nor the YARN_CONF_DIR environment variable is set. The Flink YARN Client needs one of these to be set to properly load the Hadoop configuration for accessing YARN.
2019-08-15 03:52:50,961 INFO org.apache.flink.yarn.AbstractYarnClusterDescriptor - Cluster specification: ClusterSpecification{masterMemoryMB=1024, taskManagerMemoryMB=1024, numberTaskManagers=1, slotsPerTaskManager=2}
2019-08-15 03:52:51,410 WARN org.apache.flink.yarn.AbstractYarnClusterDescriptor - The configuration directory ('/opt/module/flink-1.8.0/conf') contains both LOG4J and Logback configuration files. Please delete or rename one of them.
2019-08-15 03:52:52,456 INFO org.apache.flink.yarn.AbstractYarnClusterDescriptor - Submitting application master application_1565840709386_0002
2019-08-15 03:52:52,481 INFO org.apache.hadoop.yarn.client.api.impl.YarnClientImpl - Submitted application application_1565840709386_0002
2019-08-15 03:52:52,481 INFO org.apache.flink.yarn.AbstractYarnClusterDescriptor - Waiting for the cluster to be allocated
2019-08-15 03:52:52,484 INFO org.apache.flink.yarn.AbstractYarnClusterDescriptor - Deploying cluster, current state ACCEPTED
2019-08-15 03:52:56,776 INFO org.apache.flink.yarn.AbstractYarnClusterDescriptor - YARN application has been deployed successfully.
Starting execution of program
Printing result to stdout. Use --output to specify output path.
(abstractions,1)
(an,3)
(and,3)
(application,1)
(at,2)
(be,1)
(broadcast,2)
(called,1)
(can,1)
(deep,1)
(dive,1)
(dynamic,1)
(event,1)
(every,1)
(example,1)
(explain,1)
(exposed,1)
(flink,6)
(has,1)
(implementation,1)
(into,1)
(is,4)
(look,1)
(make,1)
(of,6)
(on,1)
(one,2)
(physical,1)
(runtime,1)
(s,3)
(stack,3)
(state,3)
(the,6)
(this,2)
(types,1)
(up,1)
(what,1)
(a,2)
(about,1)
(apache,3)
(applied,1)
(components,1)
(core,2)
(detail,1)
(evaluates,1)
(first,1)
(how,1)
(in,4)
(it,1)
(job,1)
(module,1)
(multiple,1)
(network,3)
(operator,1)
(operators,1)
(optimisations,1)
(patterns,1)
(post,2)
(posts,1)
(series,1)
(show,1)
(sitting,1)
(stream,2)
(that,2)
(their,1)
(to,2)
(various,1)
(we,2)
(which,2)
Program execution finished
Job with JobID 11307954aeb6a6356cd7b4068f0f2160 has finished.
Job Runtime: 8448 ms
Accumulator Results:
- f0f87f15adda6b1c2703a30e110db5ed (java.util.ArrayList) [69 elements]
公司:
flink run -p 2 -m yarn-cluster -yn 2 -yqu root.users.airflow -ynm opay-metrics -ys 1 -d -c com.opay.bd.opay.main.OpayOrderMetricsMain bd-flink-project-1.0.jar
flink run -p 2 -m yarn-cluster -yn 2 -yqu root.users.airflow -ynm oride-metrics -ys 1 -d -c com.opay.bd.oride.main.OrideOrderMetricsMain bd-flink-project-1.0.jar
-p,--parallelism <parallelism> 运行程序的并行度。 可以选择覆盖配置中指定的默认值
-yn 分配 YARN 容器的数量(=TaskManager 的数量)
-yqu,--yarnqueue <arg> 指定 YARN 队列
-ynm oride-metrics 给应用程序一个自定义的名字显示在 YARN 上
-ys,--yarnslots <arg> 每个 TaskManager 的槽位数量
-ys,--yarnslots <arg> 每个 TaskManager 的槽位数量
-c,--class <classname> 程序入口类
("main" 方法 或 "getPlan()" 方法)
-m yarn-cluster cluster模式