Dolphin Scheduler 1.2.1部署参数分析

Dolphin Scheduler目录配置文件解读

image.png

  • bin 启动脚本
  • conf 配置文件
  • lib ds依赖的jar包
  • script 数据库创建升级脚本，部署分发脚本
  • sql ds的元数据创建升级sql文件
  • ui 前端资源
  • install脚本 部署ds主要的配置文件修改处

bin

bin目录下比较重要的是dolphinscheduler-daemon文件，之前版本中极容易出现的找不到jdk问题来源，当前版本的jdk已经export了本机的$JAVA_HOME，再也不用担心找不到jdk了。
image.png

conf

非常重要的配置文件目录！！！非常重要的配置文件目录！！！非常重要的配置文件目录！！！
image.png

  • env目录下的dolphinscheduller_env文件中记录了所有跟ds-task相关的环境变量,1.2.1版本增加了spark版本切换功能，特别注意SPARK_HOME1和SPARK_HOME2不能注释掉，需要正确配置，实现spark组件版本切换。下面给出CDH中的配置，测试环境中没有部署Flink，请忽略Flink的配置。
  1. export HADOOP_HOME=/opt/cloudera/parcels/CDH/lib/hadoop
  2. export HADOOP_CONF_DIR=/opt/cloudera/parcels/CDH/lib/hadoop/etc/hadoop
  3. export SPARK_HOME1=/opt/cloudera/parcels/CDH/lib/spark
  4. export SPARK_HOME2=/opt/cloudera/parcels/SPARK2/lib/spark2
  5. export PYTHON_HOME=/usr/local/anaconda3/bin/python
  6. export JAVA_HOME=/usr/java/jdk1.8.0_131
  7. export HIVE_HOME=/opt/cloudera/parcels/CDH/lib/hive
  8. export FLINK_HOME=/opt/soft/flink
  9. export PATH=$HADOOP_HOME/bin:$SPARK_HOME1/bin:$SPARK_HOME2/bin:$PYTHON_HOME:$JAVA_HOME/bin:$HIVE_HOME/bin:$PATH:$FLINK_HOME/bin:$PATH
  • application.properties重要，记录了ds的元数据库配置，master和worker的配置。重要配置如下：
    • 元数据库ds默认是pg，如果需要调整为MySQL，需要在lib目录下放入MySQL的jdbc-jar包
    • 这里配置了master和worker的执行线程数量，可以根据环境进行调整
    • worker.reserved.memory是worker的内存阈值，现在默认是0.1G，已经偏小不用再次修改
  1. # base spring data source configuration
  2. spring.datasource.type=com.alibaba.druid.pool.DruidDataSource
  3. # postgre
  4. spring.datasource.driver-class-name=org.postgresql.Driver
  5. spring.datasource.url=jdbc:postgresql://localhost:5432/dolphinscheduler
  6. # mysql
  7. #spring.datasource.driver-class-name=com.mysql.jdbc.Driver
  8. #spring.datasource.url=jdbc:mysql://192.168.xx.xx:3306/dolphinscheduler?useUnicode=true&characterEncoding=UTF-8
  9. spring.datasource.username=test
  10. spring.datasource.password=test
  12. # master settings
  13. # master execute thread num
  14. master.exec.threads=100
  16. # worker settings
  17. # worker execute thread num
  18. worker.exec.threads=100
  20. # only larger than reserved memory, worker server can work. default value : physical memory * 1/6, unit is G.
  21. worker.reserved.memory=0.1
  • quartz.properties，记录了ds的quartz元数据
    • 默认依然是pg，如果需要调整为MySQL，需要在lib目录下放入MySQL的jdbc-jar包
    • 数据库选型的修改不用在这里修改，参数统一在install.sh中进行修改，这里只是给出参数的影响范围
  1. #org.quartz.jobStore.driverDelegateClass = org.quartz.impl.jdbcjobstore.StdJDBCDelegate
  2. org.quartz.jobStore.driverDelegateClass = org.quartz.impl.jdbcjobstore.PostgreSQLDelegate
  3. #org.quartz.dataSource.myDs.driver = com.mysql.jdbc.Driver
  4. org.quartz.dataSource.myDs.driver = org.postgresql.Driver
  5. #org.quartz.dataSource.myDs.URL = jdbc:mysql://192.168.xx.xx:3306/dolphinscheduler?characterEncoding=utf8
  6. org.quartz.dataSource.myDs.URL = jdbc:postgresql://localhost:5432/dolphinscheduler?characterEncoding=utf8
  7. org.quartz.dataSource.myDs.user = test
  8. org.quartz.dataSource.myDs.password = test

install脚本

install.sh部署脚本是ds部署中的重头戏，下面将参数分组进行分析。

数据库配置

  1. # for example postgresql or mysql ...
  2. dbtype="postgresql"
  4. # db config
  5. # db address and port
  6. dbhost="192.168.xx.xx:5432"
  8. # db name
  9. dbname="dolphinscheduler"
  11. # db username
  12. username="xx"
  14. # db passwprd
  15. # Note: if there are special characters, please use the \ transfer character to transfer
  16. passowrd="xx"

dbtype参数可以设置postgresql和mysql，这里指定了ds连接元数据库的jdbc相关信息

部署用户&目录

  1. # conf/config/install_config.conf config
  2. # Note: the installation path is not the same as the current path (pwd)
  3. installPath="/opt/ds-agent"
  5. # deployment user
  6. # Note: the deployment user needs to have sudo privileges and permissions to operate hdfs. If hdfs is enabled, the root directory needs to be created by itself
  7. deployUser="dolphinscheduler"

installPath是安装路径，在执行install.sh之后，会把ds安装到指定目录，如/opt/ds-agent。installPath不要和当前要一键安装的install.sh是同一目录。
deployUser是指ds的部署用户，该用户需要在部署ds的机器上打通sudo免密，并且需要具有操作hdfs的权限，建议挂到hadoop的supergroup组下。

zk集群&角色指定

配置zk集群的时候，特别注意：要用ip:2181的方式配置上去，一定要把端口带上。
ds一共包括master worker alert api四种角色，其中alert api只需指定一台机器即可，master和worker可以部署多态机器。下面的例子就是在4台机器中，部署2台master，2台worker，1台alert，1台api
ips参数，填写所有需要部署机器的hostname
masters，填写部署master机器的hostname
workers，填写部署worker机器的hostname
alertServer，填写部署alert机器的hostname
apiServers，填写部署api机器的hostname
zkroot参数可以通过调整，在一套zk集群中，托管多个ds集群，如配置zkRoot=”/dspro”,zkRoot=”/dstest”

  1. # zk cluster
  2. zkQuorum="192.168.xx.xx:2181,192.168.xx.xx:2181,192.168.xx.xx:2181"
  4. # install hosts
  5. # Note: install the scheduled hostname list. If it is pseudo-distributed, just write a pseudo-distributed hostname
  6. ips="192.168.0.1,192.168.0.2,192.168.0.3，192.168.0.4"
  8. # ssh port, default 22
  9. # Note: if ssh port is not default, modify here
  10. sshPort=22
  12. # run master machine
  13. # Note: list of hosts hostname for deploying master
  14. masters="192.168.0.1,192.168.0.2"
  16. # run worker machine
  17. # note: list of machine hostnames for deploying workers
  18. workers="192.168.0.3,192.168.0.4"
  20. # run alert machine
  21. # note: list of machine hostnames for deploying alert server
  22. alertServer="192.168.0.1"
  24. # run api machine
  25. # note: list of machine hostnames for deploying api server
  26. apiServers="192.168.0.1"
  28. # zk config
  29. # zk root directory
  30. zkRoot="/dolphinscheduler"
  32. # zk session timeout
  33. zkSessionTimeout="300"
  35. # zk connection timeout
  36. zkConnectionTimeout="300"
  38. # zk retry interval
  39. zkRetryMaxSleep="100"
  41. # zk retry maximum number of times
  42. zkRetryMaxtime="5"

邮件配置&excel文件路径

邮件配置这块也是大家非常容易出问题的，建议可以拉一下ds的代码，跑一下alert.MailUtilisTest这个测试类，下面给出QQ邮箱配置方式。如果是内网邮箱，需要注意的是ssl是否需要关闭，以及mail.user登陆用户是否需要去掉邮箱后缀。
excel路径则需要保证该路径的写入权限

  1. #QQ邮箱配置
  2. # alert config
  3. # mail protocol
  4. mailProtocol="SMTP"
  6. # mail server host
  7. mailServerHost="smtp.qq.com"
  9. # mail server port
  10. mailServerPort="465"
  12. # sender
  13. mailSender="783xx8369@qq.com"
  15. # user
  16. mailUser="783xx8369@qq.com"
  18. # sender password
  19. mailPassword="邮箱授权码"
  21. # TLS mail protocol support
  22. starttlsEnable="false"
  24. sslTrust="smtp.qq.com"
  26. # SSL mail protocol support
  27. # note: The SSL protocol is enabled by default. 
  28. # only one of TLS and SSL can be in the true state.
  29. sslEnable="true"
  31. # download excel path
  32. xlsFilePath="/tmp/xls"

资源中心&YARN

ds的资源中心支持HDFS和S3.
resUploadStartupType=”HDFS”则开启hdfs作为资源中心。
defaultFS，如果hdfs没有配置HA则需要在这里写上单点namenode的ip，如果HDFS是HA则需要将集群的core-site.xml文件和hdfs-site.xml文件拷贝到conf目录下
yarnHaIps，如果yarn启用了HA，配置两个resourcemanager的ip，如果是单点，配置空字符串
singleYarnIp，如果yarn是单点，配置resourcemanager的ip
hdfsPath，HDFS上ds存储资源的根路径，可采用默认值，如果是从1.1.0版本进行升级，需要注意这个地方，改为/escheduler

  1. # resource Center upload and select storage method：HDFS,S3,NONE
  2. resUploadStartupType="NONE"
  4. # if resUploadStartupType is HDFS，defaultFS write namenode address，HA you need to put core-site.xml and hdfs-site.xml in the conf directory.
  5. # if S3，write S3 address，HA，for example ：s3a://dolphinscheduler，
  6. # Note，s3 be sure to create the root directory /dolphinscheduler
  7. defaultFS="hdfs://mycluster:8020"
  9. # if S3 is configured, the following configuration is required.
  10. s3Endpoint="http://192.168.xx.xx:9010"
  11. s3AccessKey="xxxxxxxxxx"
  12. s3SecretKey="xxxxxxxxxx"
  14. # resourcemanager HA configuration, if it is a single resourcemanager, here is yarnHaIps=""
  15. yarnHaIps="192.168.xx.xx,192.168.xx.xx"
  17. # if it is a single resourcemanager, you only need to configure one host name. If it is resourcemanager HA, the default configuration is fine.
  18. singleYarnIp="ark1"
  20. # hdfs root path, the owner of the root path must be the deployment user. 
  21. # versions prior to 1.1.0 do not automatically create the hdfs root directory, you need to create it yourself.
  22. hdfsPath="/dolphinscheduler"
  24. # have users who create directory permissions under hdfs root path /
  25. # Note: if kerberos is enabled, hdfsRootUser="" can be used directly.
  26. hdfsRootUser="hdfs"

开发状态

devState在测试环境部署的时候可以调为true，生产环境部署建议调为false

  1. # development status, if true, for the SHELL script, you can view the encapsulated SHELL script in the execPath directory. 
  2. # If it is false, execute the direct delete
  3. devState="true"

角色参数

下面的参数主要是调整的application.properties里边的配置，涉及master,worker和apiserver
apiServerPort可以自定义修改apiserver的端口，注意需要跟前端保持一致。
master和worker的参数，初次部署建议保持默认值，如果在运行当中出现性能问题在作调整，有条件可以压一下自身环境中的master和worker的最佳线程数。

  1. # master config 
  2. # master execution thread maximum number, maximum parallelism of process instance
  3. masterExecThreads="100"
  5. # the maximum number of master task execution threads, the maximum degree of parallelism for each process instance
  6. masterExecTaskNum="20"
  8. # master heartbeat interval
  9. masterHeartbeatInterval="10"
  11. # master task submission retries
  12. masterTaskCommitRetryTimes="5"
  14. # master task submission retry interval
  15. masterTaskCommitInterval="1000"
  17. # master maximum cpu average load, used to determine whether the master has execution capability
  18. masterMaxCpuLoadAvg="100"
  20. # master reserve memory to determine if the master has execution capability
  21. masterReservedMemory="0.1"
  23. # worker config
  24. # worker execution thread
  25. workerExecThreads="100"
  27. # worker heartbeat interval
  28. workerHeartbeatInterval="10"
  30. # worker number of fetch tasks
  31. workerFetchTaskNum="3"
  33. # worker reserve memory to determine if the master has execution capability
  34. workerReservedMemory="0.1"
  36. # api config
  37. # api server port
  38. apiServerPort="12345"

特别注意

  • ds需要启用资源中心之后，才可以创建租户，因此资源中心的配置一定要正确
  • ds老版本部署需要配置JDK的问题已经解决
  • installPath不要和当前要一键安装的install.sh是同一目录
  • ds的task运行都依赖env目录下的环境变量文件，需要正确配置
  • HDFS高可用，需要把core-site.xml和hdfs-site.xml文件拷贝到conf目录下
  • 邮件配置中mailUser和mailSender的区别