docker下的spark集群，调整参数榨干硬件
Posted 程序员欣宸
tags:
篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了docker下的spark集群，调整参数榨干硬件相关的知识，希望对你有一定的参考价值。
欢迎访问我的GitHub
本篇概览
-
本文是《docker下，极速搭建spark集群(含hdfs集群)》的续篇，前文将spark集群搭建成功并进行了简单的验证，但是存在以下几个小问题：
- spark只有一个work节点，只适合处理小数据量的任务，遇到大量数据的任务要消耗更多时间；
- hdfs的文件目录和docker安装目录在一起，如果要保存大量文件，很可能由于磁盘空间不足导致上传失败；
- master的4040和work的8080端口都没有开放，看不到job、stage、executor的运行情况；
- 今天就来调整系统参数，解决上述问题；
最初的docker-compose.yml内容
- 优化前的docker-compose.yml内容如下所示：
version: "2.2"
services:
namenode:
image: bde2020/hadoop-namenode:1.1.0-hadoop2.7.1-java8
container_name: namenode
volumes:
- hadoop_namenode:/hadoop/dfs/name
- ./input_files:/input_files
environment:
- CLUSTER_NAME=test
env_file:
- ./hadoop.env
ports:
- 50070:50070
resourcemanager:
image: bde2020/hadoop-resourcemanager:1.1.0-hadoop2.7.1-java8
container_name: resourcemanager
depends_on:
- namenode
- datanode1
- datanode2
env_file:
- ./hadoop.env
historyserver:
image: bde2020/hadoop-historyserver:1.1.0-hadoop2.7.1-java8
container_name: historyserver
depends_on:
- namenode
- datanode1
- datanode2
volumes:
- hadoop_historyserver:/hadoop/yarn/timeline
env_file:
- ./hadoop.env
nodemanager1:
image: bde2020/hadoop-nodemanager:1.1.0-hadoop2.7.1-java8
container_name: nodemanager1
depends_on:
- namenode
- datanode1
- datanode2
env_file:
- ./hadoop.env
datanode1:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode1
depends_on:
- namenode
volumes:
- hadoop_datanode1:/hadoop/dfs/data
env_file:
- ./hadoop.env
datanode2:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode2
depends_on:
- namenode
volumes:
- hadoop_datanode2:/hadoop/dfs/data
env_file:
- ./hadoop.env
datanode3:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode3
depends_on:
- namenode
volumes:
- hadoop_datanode3:/hadoop/dfs/data
env_file:
- ./hadoop.env
master:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: master
command: bin/spark-class org.apache.spark.deploy.master.Master -h master
hostname: master
environment:
MASTER: spark://master:7077
SPARK_CONF_DIR: /conf
SPARK_PUBLIC_DNS: localhost
links:
- namenode
expose:
- 7001
- 7002
- 7003
- 7004
- 7005
- 7077
- 6066
ports:
- 6066:6066
- 7077:7077
- 8080:8080
volumes:
- ./conf/master:/conf
- ./data:/tmp/data
- ./jars:/root/jars
worker:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 1g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8081
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
ports:
- 8081:8081
volumes:
- ./conf/worker:/conf
- ./data:/tmp/data
volumes:
hadoop_namenode:
hadoop_datanode1:
hadoop_datanode2:
hadoop_datanode3:
hadoop_historyserver:- 接下来开始优化；
实战环境信息
- 本次实战所用的电脑是联想笔记本：
- CPU：i5-6300HQ（四核四线程）
- 内存：16G
- 硬盘：256G的NVMe再加500G机械硬盘
- 系统：Deepin15
- docker：18.09.1
- docker-compose：1.17.1
- spark：2.3.0
- hdfs：2.7.1
调整work节点数量
- 由于内存有16G，于是打算将work节点数从1个调整到6个，调整后work容器的配置如下：
worker1:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker1
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker1
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8081
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
volumes:
- ./conf/worker1:/conf
- ./data/worker1:/tmp/data
worker2:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker2
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker2
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8082
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
volumes:
- ./conf/worker2:/conf
- ./data/worker2:/tmp/data- 如上所示，注意volumes参数，都映射在了docker-compose.yml同一层级的conf和data两个目录下，这里只贴出了worker1和worker2的内容，worker3-worker6的内容都是类似的；
hdfs的文件目录导致的磁盘空间不足问题
- 先来看下hdfs的文件目录配置：
volumes:
- hadoop_datanode1:/hadoop/dfs/data- 上面的hadoop_datanode1数据卷的配置在docker-compose.yml的最底部，是默认声明，如下：
volumes:
hadoop_namenode:
hadoop_datanode1:
hadoop_datanode2:
hadoop_datanode3:
hadoop_historyserver:- 在容器运行状态，执行命令docker inspect datanode1查看容器信息，和数据卷相关的信息如下所示：
"Mounts": [
"Type": "volume",
"Name": "temp_hadoop_datanode1",
"Source": "/var/lib/docker/volumes/temp_hadoop_datanode1/_data",
"Destination": "/hadoop/dfs/data",
"Driver": "local",
"Mode": "rw",
"RW": true,
"Propagation": ""
]- 可见hdfs容器的文件目录对应的是宿主机的/var/lib/docker/volumes；
- 用df -m看看磁盘空间情况，如下所示，"/var/lib/docker/volumes"所在的"/dev/nvme0n1p3"设备可用空间只有20多G（29561），显然在保存大量文件时这个空间是不够的，而且hdfs的默认副本数为3：
root@willzhao-deepin:/data/work/spark/temp# df -m
文件系统 1M-块 已用 可用 已用% 挂载点
udev 7893 0 7893 0% /dev
tmpfs 1584 4 1581 1% /run
/dev/nvme0n1p3 43927 12107 29561 30% /
tmpfs 7918 0 7918 0% /dev/shm
tmpfs 5 1 5 1% /run/lock
tmpfs 7918 0 7918 0% /sys/fs/cgroup
/dev/nvme0n1p4 87854 181 83169 1% /home
/dev/nvme0n1p1 300 7 293 3% /boot/efi
/dev/sda1 468428 109152 335430 25% /data
tmpfs 1584 1 1584 1% /run/user/108
tmpfs 1584 0 1584 0% /run/user/0- 上面的磁盘信息显示设备/dev/sda1还有300G，所以hdfs的文件目录映射到/dev/sda1就能缓解磁盘空间问题了，于是修改docker-compose.yml文件中hdfs的三个数据节点的配置，修改后如下：
datanode1:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode1
depends_on:
- namenode
volumes:
- ./hadoop/datanode1:/hadoop/dfs/data
env_file:
- ./hadoop.env
datanode2:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode2
depends_on:
- namenode
volumes:
- ./hadoop/datanode2:/hadoop/dfs/data
env_file:
- ./hadoop.env
datanode3:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode3
depends_on:
- namenode
volumes:
- ./hadoop/datanode3:/hadoop/dfs/data
env_file:
- ./hadoop.env- 再将下面这段配置删除：
volumes:
hadoop_namenode:
hadoop_datanode1:
hadoop_datanode2:
hadoop_datanode3:
hadoop_historyserver:开发master的4040和work的8080端口
- 任务运行过程中，如果有UI页面来观察详情，可以帮助我们更全面直观的了解运行情况，所以需要修改配置开放端口；
- 如下所示，expose参数增加4040，表示对外暴露4040端口，ports参数增加4040:4040，表示容器的4040映射到宿主机的4040端口：
master:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: master
command: bin/spark-class org.apache.spark.deploy.master.Master -h master
hostname: master
environment:
MASTER: spark://master:7077
SPARK_CONF_DIR: /conf
SPARK_PUBLIC_DNS: localhost
links:
- namenode
expose:
- 4040
- 7001
- 7002
- 7003
- 7004
- 7005
- 7077
- 6066
ports:
- 4040:4040
- 6066:6066
- 7077:7077
- 8080:8080
volumes:
- ./conf/master:/conf
- ./data:/tmp/data
- ./jars:/root/jars- worker的web端口同样需要打开，访问worker的web页面可以观察worker的状态，并且查看任务日志（这个很重要），这里要注意的是由于有多个worker，所以要映射到宿主机的多个端口，如下配置，workder1的environment.SPARK_WORKER_WEBUI_PORT配置为8081，并且暴露8081，再将容器的8081映射到宿主机的8081，workder2的environment.SPARK_WORKER_WEBUI_PORT配置为8082，并且暴露8082，再将容器的8082映射到宿主机的8082：
worker1:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker1
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker1
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8081
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8081
ports:
- 8081:8081
volumes:
- ./conf/worker1:/conf
- ./data/worker1:/tmp/data
worker2:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker2
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker2
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8082
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8082
ports:
- 8082:8082
volumes:
- ./conf/worker2:/conf
- ./data/worker2:/tmp/data -
worker3-worker6的配置与上面类似，注意用不同的端口号；
- 至此，修改已经完成，最终版的docker-compose.yml内容如下：
version: "2.2"
services:
namenode:
image: bde2020/hadoop-namenode:1.1.0-hadoop2.7.1-java8
container_name: namenode
volumes:
- ./hadoop/namenode:/hadoop/dfs/name
- ./input_files:/input_files
environment:
- CLUSTER_NAME=test
env_file:
- ./hadoop.env
ports:
- 50070:50070
resourcemanager:
image: bde2020/hadoop-resourcemanager:1.1.0-hadoop2.7.1-java8
container_name: resourcemanager
depends_on:
- namenode
- datanode1
- datanode2
env_file:
- ./hadoop.env
historyserver:
image: bde2020/hadoop-historyserver:1.1.0-hadoop2.7.1-java8
container_name: historyserver
depends_on:
- namenode
- datanode1
- datanode2
volumes:
- ./hadoop/historyserver:/hadoop/yarn/timeline
env_file:
- ./hadoop.env
nodemanager1:
image: bde2020/hadoop-nodemanager:1.1.0-hadoop2.7.1-java8
container_name: nodemanager1
depends_on:
- namenode
- datanode1
- datanode2
env_file:
- ./hadoop.env
datanode1:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode1
depends_on:
- namenode
volumes:
- ./hadoop/datanode1:/hadoop/dfs/data
env_file:
- ./hadoop.env
datanode2:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode2
depends_on:
- namenode
volumes:
- ./hadoop/datanode2:/hadoop/dfs/data
env_file:
- ./hadoop.env
datanode3:
image: bde2020/hadoop-datanode:1.1.0-hadoop2.7.1-java8
container_name: datanode3
depends_on:
- namenode
volumes:
- ./hadoop/datanode3:/hadoop/dfs/data
env_file:
- ./hadoop.env
master:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: master
command: bin/spark-class org.apache.spark.deploy.master.Master -h master
hostname: master
environment:
MASTER: spark://master:7077
SPARK_CONF_DIR: /conf
SPARK_PUBLIC_DNS: localhost
links:
- namenode
expose:
- 4040
- 7001
- 7002
- 7003
- 7004
- 7005
- 7077
- 6066
ports:
- 4040:4040
- 6066:6066
- 7077:7077
- 8080:8080
volumes:
- ./conf/master:/conf
- ./data:/tmp/data
- ./jars:/root/jars
worker1:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker1
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker1
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8081
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8081
ports:
- 8081:8081
volumes:
- ./conf/worker1:/conf
- ./data/worker1:/tmp/data
worker2:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker2
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker2
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8082
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8082
ports:
- 8082:8082
volumes:
- ./conf/worker2:/conf
- ./data/worker2:/tmp/data
worker3:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker3
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker3
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8083
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8083
ports:
- 8083:8083
volumes:
- ./conf/worker3:/conf
- ./data/worker3:/tmp/data
worker4:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker4
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker4
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8084
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8084
ports:
- 8084:8084
volumes:
- ./conf/worker4:/conf
- ./data/worker4:/tmp/data
worker5:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker5
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker5
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8085
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8085
ports:
- 8085:8085
volumes:
- ./conf/worker5:/conf
- ./data/worker5:/tmp/data
worker6:
image: gettyimages/spark:2.3.0-hadoop-2.8
container_name: worker6
command: bin/spark-class org.apache.spark.deploy.worker.Worker spark://master:7077
hostname: worker6
environment:
SPARK_CONF_DIR: /conf
SPARK_WORKER_CORES: 2
SPARK_WORKER_MEMORY: 2g
SPARK_WORKER_PORT: 8881
SPARK_WORKER_WEBUI_PORT: 8086
SPARK_PUBLIC_DNS: localhost
links:
- master
expose:
- 7012
- 7013
- 7014
- 7015
- 8881
- 8086
ports:
- 8086:8086
volumes:
- ./conf/worker6:/conf
- ./data/worker6:/tmp/data- 接下来我们运行一个实例来验证；
验证
- 在docker-compose.yml所在目录创建hadoop.env文件，内容如下：
CORE_CONF_fs_defaultFS=hdfs://namenode:8020
CORE_CONF_hadoop_http_staticuser_user=root
CORE_CONF_hadoop_proxyuser_hue_hosts=*
CORE_CONF_hadoop_proxyuser_hue_groups=*
HDFS_CONF_dfs_webhdfs_enabled=true
HDFS_CONF_dfs_permissions_enabled=false
YARN_CONF_yarn_log___aggregation___enable=true
YARN_CONF_yarn_resourcemanager_recovery_enabled=true
YARN_CONF_yarn_resourcemanager_store_class=org.apache.hadoop.yarn.server.resourcemanager.recovery.FileSystemRMStateStore
YARN_CONF_yarn_resourcemanager_fs_state___store_uri=/rmstate
YARN_CONF_yarn_nodemanager_remote___app___log___dir=/app-logs
YARN_CONF_yarn_log_server_url=http://historyserver:8188/applicationhistory/logs/
YARN_CONF_yarn_timeline___service_enabled=true
YARN_CONF_yarn_timeline___service_generic___application___history_enabled=true
YARN_CONF_yarn_resourcemanager_system___metrics___publisher_enabled=true
YARN_CONF_yarn_resourcemanager_hostname=resourcemanager
YARN_CONF_yarn_timeline___service_hostname=historyserver
YARN_CONF_yarn_resourcemanager_address=resourcemanager:8032
YARN_CONF_yarn_resourcemanager_scheduler_address=resourcemanager:8030
YARN_CONF_yarn_resourcemanager_resource___tracker_address=resourcemanager:8031- 修改好docker-composes.yml后，执行以下命令启动容器：
docker-compose up -d-
此次验证所用的spark应用的功能是分析维基百科的网站统计信息，找出访问量最大的网页，本次实战用现成的jar包，不涉及编码，该应用的源码和开发详情请参照《spark实战之：分析维基百科网站统计数据(java版)》；
- 从github下载已经构建好的spark应用jar文件：
wget https://raw.githubusercontent.com/zq2599/blog_demos/master/files/sparkdemo-1.0-SNAPSHOT.jar- 从github下载维基百科的网站统计信息大数据集，这里只下载了一个文件，建议您参照《寻找海量数据集用于大数据开发实战(维基百科网站统计数据)》下载更多文件用来实战：
wget https://raw.githubusercontent.com/zq2599/blog_demos/master/files/pagecounts-20160801-000000- 将下载的sparkdemo-1.0-SNAPSHOT.jar文件放在docker-compose.xml所在目录的jars目录下；
- 在docker-compose.xml所在目录的input_files目录内创建input目录，再将下载的pagecounts-20160801-000000文件放在这个input目录下；
- 执行以下命令，将整个input目录放入hdfs：
docker exec namenode hdfs dfs -put /input_files/input /- 执行以下命令，提交一个任务，使用了12个executor，每个1G内存：
docker exec -it master spark-submit \\
--class com.bolingcavalry.sparkdemo.app.WikiRank \\
--executor-memory 1g \\
--total-executor-cores 12 \\
/root/jars/sparkdemo-1.0-SNAPSHOT.jar \\
namenode \\
8020-
宿主机的状态如下所示，CPU和内存都被榨干：
-
宿主机的IP地址是192.168.1.102，以下是状态信息，地址：http://192.168.1.102:8080/
-
查看job的Stage情况，如下图，这些信息对学习和掌握spark至关重要，地址：http://192.168.1.102:4040
-
查看worker1的基本情况，如下图，地址是：http://192.168.1.102:8081
-
如果想查看worker1上的业务日志，请点击下图红框中的链接，但此时会提示页面访问失败，对应的url是"http://localhost:8081/logPage?appId=app-20190216081637-0002&executorId=5&logType=stdout"，这个地址是页面生成的，我们只要把其中的"localhost"改成宿主机的IP地址就好了：
-
修改后的链接可以访问，看到的业务日志如下图，红框中就是业务代码输出的日志：
-
以上就是优化和验证的全部过程，您可以根据自己机器的实际情况来调整参数，将电脑的性能充分的利用起来；
- 后来我用24个300M的文件做数据集，大约1.5亿条记录，在上述硬件环境运行上述命令，最终耗时30分钟完成，如下图：
欢迎关注51CTO博客：程序员欣宸
以上是关于docker下的spark集群，调整参数榨干硬件的主要内容，如果未能解决你的问题，请参考以下文章