Browse Source

fix bug, update action would change the user of definition.

pull/1/MERGE
baoliang 6 years ago
parent
commit
cba13a1fee
  1. 431
      docs/zh_CN/后端部署文档.md
  2. 2
      install.sh

431
docs/zh_CN/后端部署文档.md

@ -86,370 +86,11 @@ escheduler ALL=(ALL) NOPASSWD: NOPASSWD: ALL
#Default requiretty
```
## 配置文件说明
```
说明:配置文件位于 target/escheduler-{version}/conf 下面
```
### escheduler-alert
配置邮件告警信息
* alert.properties
```
#以qq邮箱为例,如果是别的邮箱,请更改对应配置
#alert type is EMAIL/SMS
alert.type=EMAIL
# mail server configuration
mail.protocol=SMTP
mail.server.host=smtp.exmail.qq.com
mail.server.port=25
mail.sender=xxxxxxx@qq.com
mail.passwd=xxxxxxx
# xls file path, need manually create it before use if not exist
xls.file.path=/opt/xls
```
### escheduler-common
通用配置文件配置,队列选择及地址配置,通用文件目录配置
- common/common.properties
```
#task queue implementation, default "zookeeper"
escheduler.queue.impl=zookeeper
# user data directory path, self configuration, please make sure the directory exists and have read write permissions
data.basedir.path=/tmp/escheduler
# directory path for user data download. self configuration, please make sure the directory exists and have read write permissions
data.download.basedir.path=/tmp/escheduler/download
# process execute directory. self configuration, please make sure the directory exists and have read write permissions
process.exec.basepath=/tmp/escheduler/exec
# data base dir, resource file will store to this hadoop hdfs path, self configuration, please make sure the directory exists on hdfs and have read write permissions。"/escheduler" is recommended
data.store2hdfs.basepath=/escheduler
# whether hdfs starts
hdfs.startup.state=true
# system env path. self configuration, please make sure the directory and file exists and have read write execute permissions
escheduler.env.path=/opt/.escheduler_env.sh
escheduler.env.py=/opt/escheduler_env.py
#resource.view.suffixs
resource.view.suffixs=txt,log,sh,conf,cfg,py,java,sql,hql,xml
# is development state? default "false"
development.state=false
```
SHELL任务 环境变量配置
```
说明:配置文件位于 target/escheduler-{version}/conf/env 下面,这个会是Worker执行任务时加载的环境
```
.escheduler_env.sh
```
export HADOOP_HOME=/opt/soft/hadoop
export HADOOP_CONF_DIR=/opt/soft/hadoop/etc/hadoop
export SPARK_HOME1=/opt/soft/spark1
export SPARK_HOME2=/opt/soft/spark2
export PYTHON_HOME=/opt/soft/python
export JAVA_HOME=/opt/soft/java
export HIVE_HOME=/opt/soft/hive
export PATH=$HADOOP_HOME/bin:$SPARK_HOME1/bin:$SPARK_HOME2/bin:$PYTHON_HOME/bin:$JAVA_HOME/bin:$HIVE_HOME/bin:$PATH
```
Python任务 环境变量配置
```
说明:配置文件位于 target/escheduler-{version}/conf/env 下面
```
escheduler_env.py
```
import os
HADOOP_HOME="/opt/soft/hadoop"
SPARK_HOME1="/opt/soft/spark1"
SPARK_HOME2="/opt/soft/spark2"
PYTHON_HOME="/opt/soft/python"
JAVA_HOME="/opt/soft/java"
HIVE_HOME="/opt/soft/hive"
PATH=os.environ['PATH']
PATH="%s/bin:%s/bin:%s/bin:%s/bin:%s/bin:%s/bin:%s"%(HIVE_HOME,HADOOP_HOME,SPARK_HOME1,SPARK_HOME2,JAVA_HOME,PYTHON_HOME,PATH)
os.putenv('PATH','%s'%PATH)
```
hadoop 配置文件
- common/hadoop/hadoop.properties
```
# ha or single namenode,If namenode ha needs to copy core-site.xml and hdfs-site.xml to the conf directory
fs.defaultFS=hdfs://mycluster:8020
#resourcemanager ha note this need ips , this empty if single
yarn.resourcemanager.ha.rm.ids=192.168.xx.xx,192.168.xx.xx
# If it is a single resourcemanager, you only need to configure one host name. If it is resourcemanager HA, the default configuration is fine
yarn.application.status.address=http://ark1:8088/ws/v1/cluster/apps/%s
```
定时器配置文件
- quartz.properties
```
#============================================================================
# Configure Main Scheduler Properties
#============================================================================
org.quartz.scheduler.instanceName = EasyScheduler
org.quartz.scheduler.instanceId = AUTO
org.quartz.scheduler.makeSchedulerThreadDaemon = true
org.quartz.jobStore.useProperties = false
#============================================================================
# Configure ThreadPool
#============================================================================
org.quartz.threadPool.class = org.quartz.simpl.SimpleThreadPool
org.quartz.threadPool.makeThreadsDaemons = true
org.quartz.threadPool.threadCount = 25
org.quartz.threadPool.threadPriority = 5
#============================================================================
# Configure JobStore
#============================================================================
org.quartz.jobStore.class = org.quartz.impl.jdbcjobstore.JobStoreTX
org.quartz.jobStore.driverDelegateClass = org.quartz.impl.jdbcjobstore.StdJDBCDelegate
org.quartz.jobStore.tablePrefix = QRTZ_
org.quartz.jobStore.isClustered = true
org.quartz.jobStore.misfireThreshold = 60000
org.quartz.jobStore.clusterCheckinInterval = 5000
org.quartz.jobStore.dataSource = myDs
#============================================================================
# Configure Datasources
#============================================================================
org.quartz.dataSource.myDs.driver = com.mysql.jdbc.Driver
org.quartz.dataSource.myDs.URL = jdbc:mysql://192.168.xx.xx:3306/escheduler?characterEncoding=utf8&useSSL=false
org.quartz.dataSource.myDs.user = xx
org.quartz.dataSource.myDs.password = xx
org.quartz.dataSource.myDs.maxConnections = 10
org.quartz.dataSource.myDs.validationQuery = select 1
```
zookeeper 配置文件
- zookeeper.properties
```
#zookeeper cluster
zookeeper.quorum=192.168.xx.xx:2181,192.168.xx.xx:2181,192.168.xx.xx:2181
#escheduler root directory
zookeeper.escheduler.root=/escheduler
#zookeeper server dirctory
zookeeper.escheduler.dead.servers=/escheduler/dead-servers
zookeeper.escheduler.masters=/escheduler/masters
zookeeper.escheduler.workers=/escheduler/workers
#zookeeper lock dirctory
zookeeper.escheduler.lock.masters=/escheduler/lock/masters
zookeeper.escheduler.lock.workers=/escheduler/lock/workers
#escheduler failover directory
zookeeper.escheduler.lock.masters.failover=/escheduler/lock/failover/masters
zookeeper.escheduler.lock.workers.failover=/escheduler/lock/failover/workers
#escheduler failover directory
zookeeper.session.timeout=300
zookeeper.connection.timeout=300
zookeeper.retry.sleep=1000
zookeeper.retry.maxtime=5
```
### escheduler-dao
dao数据源配置
- dao/data_source.properties
```
# base spring data source configuration
spring.datasource.type=com.alibaba.druid.pool.DruidDataSource
spring.datasource.driver-class-name=com.mysql.jdbc.Driver
spring.datasource.url=jdbc:mysql://192.168.xx.xx:3306/escheduler?characterEncoding=UTF-8
spring.datasource.username=xx
spring.datasource.password=xx
# connection configuration
spring.datasource.initialSize=5
# min connection number
spring.datasource.minIdle=5
# max connection number
spring.datasource.maxActive=50
# max wait time for get a connection in milliseconds. if configuring maxWait, fair locks are enabled by default and concurrency efficiency decreases.
# If necessary, unfair locks can be used by configuring the useUnfairLock attribute to true.
spring.datasource.maxWait=60000
# milliseconds for check to close free connections
spring.datasource.timeBetweenEvictionRunsMillis=60000
# the Destroy thread detects the connection interval and closes the physical connection in milliseconds if the connection idle time is greater than or equal to minEvictableIdleTimeMillis.
spring.datasource.timeBetweenConnectErrorMillis=60000
# the longest time a connection remains idle without being evicted, in milliseconds
spring.datasource.minEvictableIdleTimeMillis=300000
#the SQL used to check whether the connection is valid requires a query statement. If validation Query is null, testOnBorrow, testOnReturn, and testWhileIdle will not work.
spring.datasource.validationQuery=SELECT 1
#check whether the connection is valid for timeout, in seconds
spring.datasource.validationQueryTimeout=3
# when applying for a connection, if it is detected that the connection is idle longer than time Between Eviction Runs Millis,
# validation Query is performed to check whether the connection is valid
spring.datasource.testWhileIdle=true
#execute validation to check if the connection is valid when applying for a connection
spring.datasource.testOnBorrow=true
#execute validation to check if the connection is valid when the connection is returned
spring.datasource.testOnReturn=false
spring.datasource.defaultAutoCommit=true
spring.datasource.keepAlive=true
# open PSCache, specify count PSCache for every connection
spring.datasource.poolPreparedStatements=true
spring.datasource.maxPoolPreparedStatementPerConnectionSize=20
```
### escheduler-server
master配置文件
- master.properties
```
# master execute thread num
master.exec.threads=100
# master execute task number in parallel
master.exec.task.number=20
# master heartbeat interval
master.heartbeat.interval=10
# master commit task retry times
master.task.commit.retryTimes=5
# master commit task interval
master.task.commit.interval=100
# only less than cpu avg load, master server can work. default value : the number of cpu cores * 2
master.max.cpuload.avg=10
# only larger than reserved memory, master server can work. default value : physical memory * 1/10, unit is G.
master.reserved.memory=1
```
worker配置文件
- worker.properties
```
# worker execute thread num
worker.exec.threads=100
# worker heartbeat interval
worker.heartbeat.interval=10
# submit the number of tasks at a time
worker.fetch.task.num = 10
# only less than cpu avg load, worker server can work. default value : the number of cpu cores * 2
worker.max.cpuload.avg=10
# only larger than reserved memory, worker server can work. default value : physical memory * 1/6, unit is G.
worker.reserved.memory=1
```
### escheduler-api
web配置文件
- application.properties
```
# server port
server.port=12345
# session config
server.session.timeout=7200
server.context-path=/escheduler/
# file size limit for upload
spring.http.multipart.max-file-size=1024MB
spring.http.multipart.max-request-size=1024MB
# post content
server.max-http-post-size=5000000
```
## 伪分布式部署
### 1,创建部署用户
如上 **创建部署用户**
### 2,根据实际需求来创建HDFS根路径
根据 **common/common.properties****hdfs.startup.state** 的配置来判断是否启动HDFS,如果启动,则需要创建HDFS根路径,并将 **owner** 修改为**部署用户**,否则忽略此步骤
根据 **common/common.properties****hdf.startup.states** 的配置来判断是否启动HDFS,如果启动,则需要创建HDFS根路径,并将 **owner** 修改为**部署用户**,否则忽略此步骤
### 3,项目编译
@ -465,40 +106,6 @@ server.max-http-post-size=5000000
- 将**.escheduler_env.sh** 和 **escheduler_env.py** 两个环境变量文件复制到 **common/common.properties**配置的**escheduler.env.path****escheduler.env.py** 的目录下,并将 **owner** 修改为**部署用户**
### 6,启停服务
* 启停Master
```启动master
sh ./bin/escheduler-daemon.sh start master-server
sh ./bin/escheduler-daemon.sh stop master-server
```
* 启停Worker
```
sh ./bin/escheduler-daemon.sh start worker-server
sh ./bin/escheduler-daemon.sh stop worker-server
```
* 启停Api
```
sh ./bin/escheduler-daemon.sh start api-server
sh ./bin/escheduler-daemon.sh stop api-server
```
* 启停Logger
```
sh ./bin/escheduler-daemon.sh start logger-server
sh ./bin/escheduler-daemon.sh stop logger-server
```
* 启停Alert
```
sh ./bin/escheduler-daemon.sh start alert-server
sh ./bin/escheduler-daemon.sh stop alert-server
```
@ -546,6 +153,42 @@ sh ./bin/escheduler-daemon.sh stop alert-server
- 注意:scp_hosts.sh 里 `tar -zxvf $workDir/../escheduler-1.0.0.tar.gz -C $installPath` 中的版本号(1.0.0)需要执行前手动替换成对应的版本号
### 7,启停服务
* 启停Master
```启动master
sh ./bin/escheduler-daemon.sh start master-server
sh ./bin/escheduler-daemon.sh stop master-server
```
* 启停Worker
```
sh ./bin/escheduler-daemon.sh start worker-server
sh ./bin/escheduler-daemon.sh stop worker-server
```
* 启停Api
```
sh ./bin/escheduler-daemon.sh start api-server
sh ./bin/escheduler-daemon.sh stop api-server
```
* 启停Logger
```
sh ./bin/escheduler-daemon.sh start logger-server
sh ./bin/escheduler-daemon.sh stop logger-server
```
* 启停Alert
```
sh ./bin/escheduler-daemon.sh start alert-server
sh ./bin/escheduler-daemon.sh stop alert-server
```
## 服务监控
monitor_server.py 脚本是监听,master和worker服务挂掉重启的脚本

2
install.sh

@ -211,7 +211,7 @@ mailPassword="xxxxxxxxxx"
xlsFilePath="/opt/xls"
# conf/config/install_config.conf配置
# 安装路径
# 安装路径,不要当前路径(pwd)一样
installPath="/data1_1T/escheduler"
# 部署用户

Loading…
Cancel
Save