Pseudo-Cluster Deployment

The purpose of the pseudo-cluster deployment is to deploy the DolphinScheduler service on a single machine. In this mode, DolphinScheduler's master, worker, API server, are all on the same machine.

If you are a new hand and want to experience DolphinScheduler functions, we recommend you install follow Standalone deployment. If you want to experience more complete functions and schedule massive tasks, we recommend you install follow[pseudo-cluster deployment. If you want to deploy DolphinScheduler in production, we recommend you follow cluster deployment or Kubernetes deployment.

Preparation

Pseudo-cluster deployment of DolphinScheduler requires external software support:

  • JDK:Download JDK (1.8+), and configure JAVA_HOME to and PATH variable. You can skip this step, if it already exists in your environment.
  • Binary package: Download the DolphinScheduler binary package at download page
  • Database: PostgreSQL (8.2.15+) or MySQL (5.7+), you can choose one of the two, such as MySQL requires JDBC Driver 8.0.16
  • Registry Center: ZooKeeper (3.4.6+),download link
  • Process tree analysis
    • pstree for macOS
    • psmisc for Fedora/Red/Hat/CentOS/Ubuntu/Debian

Note: DolphinScheduler itself does not depend on Hadoop, Hive, Spark, but if you need to run tasks that depend on them, you need to have the corresponding environment support.

DolphinScheduler Startup Environment

Configure User Exemption and Permissions

Create a deployment user, and make sure to configure sudo without password. Here make an example to create user dolphinscheduler:

# To create a user, login as root
useradd dolphinscheduler

# Add password
echo "dolphinscheduler" | passwd --stdin dolphinscheduler

# Configure sudo without password
sed -i '$adolphinscheduler  ALL=(ALL)  NOPASSWD: NOPASSWD: ALL' /etc/sudoers
sed -i 's/Defaults    requirett/#Defaults    requirett/g' /etc/sudoers

# Modify directory permissions and grant permissions for user you created above
chown -R dolphinscheduler:dolphinscheduler apache-dolphinscheduler-*-bin

NOTICE:

  • Due to DolphinScheduler's multi-tenant task switch user using command sudo -u {linux-user}, the deployment user needs to have sudo privileges and be password-free. If novice learners don’t understand, you can ignore this point for now.
  • If you find the line "Defaults requirett" in the /etc/sudoers file, please comment the content.

Configure Machine SSH Password-Free Login

Since resources need to be sent to different machines during installation, SSH password-free login is required between each machine. The following shows the steps to configure password-free login:

su dolphinscheduler

ssh-keygen -t rsa -P '' -f ~/.ssh/id_rsa
cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys
chmod 600 ~/.ssh/authorized_keys

Notice: After the configuration is complete, you can run the command ssh localhost to test works or not. If you can login with ssh without password stands for successful.

Start ZooKeeper

Go to the ZooKeeper installation directory, copy configure file zoo_sample.cfg to conf/zoo.cfg, and change value of dataDir in conf/zoo.cfg to dataDir=./tmp/zookeeper.

# Start ZooKeeper
./bin/zkServer.sh start

Modify Configuration

After completing the preparation of the basic environment, you need to modify the configuration file according to the environment you used. The configuration files are both in directory bin/env and named install_env.sh and dolphinscheduler_env.sh.

Modify install_env.sh

File install_env.sh describes which machines will be installed DolphinScheduler and what server will be installed on each machine. You could find this file in the path bin/env/install_env.sh and the detail of the configuration as below.

# ---------------------------------------------------------
# INSTALL MACHINE
# ---------------------------------------------------------
# Due to the master, worker, and API server being deployed on a single node, the IP of the server is the machine IP or localhost
ips="localhost"
masters="localhost"
workers="localhost:default"
alertServer="localhost"
apiServers="localhost"

# DolphinScheduler installation path, it will auto-create if not exists
installPath="~/dolphinscheduler"

# Deploy user, use the user you create in section **Configure machine SSH password-free login**
deployUser="dolphinscheduler"

Modify dolphinscheduler_env.sh

File ./bin/env/dolphinscheduler_env.sh describes the following configurations:

  • Database configuration of DolphinScheduler, see Initialize the Database for detailed instructions.
  • Some tasks which need external dependencies or libraries such as JAVA_HOME and SPARK_HOME.
  • Registry center zookeeper.
  • Server related configuration, such as cache type, timezone, etc.

You could ignore the task external dependencies if you do not use those tasks, but you have to change JAVA_HOME, registry center and database related configurations based on your environment.

# JAVA_HOME, will use it to start DolphinScheduler server
export JAVA_HOME=${JAVA_HOME:-/opt/soft/java}

# Database related configuration, set database type, username and password
export DATABASE=${DATABASE:-postgresql}
export SPRING_PROFILES_ACTIVE=${DATABASE}
export SPRING_DATASOURCE_DRIVER_CLASS_NAME=org.postgresql.Driver
export SPRING_DATASOURCE_URL="jdbc:postgresql://127.0.0.1:5432/dolphinscheduler"
export SPRING_DATASOURCE_USERNAME={user}
export SPRING_DATASOURCE_PASSWORD={password}

# DolphinScheduler server related configuration
export SPRING_CACHE_TYPE=${SPRING_CACHE_TYPE:-none}
export SPRING_JACKSON_TIME_ZONE=${SPRING_JACKSON_TIME_ZONE:-UTC}
export MASTER_FETCH_COMMAND_NUM=${MASTER_FETCH_COMMAND_NUM:-10}

# Registry center configuration, determines the type and link of the registry center
export REGISTRY_TYPE=${REGISTRY_TYPE:-zookeeper}
export REGISTRY_ZOOKEEPER_CONNECT_STRING=${REGISTRY_ZOOKEEPER_CONNECT_STRING:-localhost:2181}

# Tasks related configurations, need to change the configuration if you use the related tasks.
export HADOOP_HOME=${HADOOP_HOME:-/opt/soft/hadoop}
export HADOOP_CONF_DIR=${HADOOP_CONF_DIR:-/opt/soft/hadoop/etc/hadoop}
export SPARK_HOME1=${SPARK_HOME1:-/opt/soft/spark1}
export SPARK_HOME2=${SPARK_HOME2:-/opt/soft/spark2}
export PYTHON_HOME=${PYTHON_HOME:-/opt/soft/python}
export HIVE_HOME=${HIVE_HOME:-/opt/soft/hive}
export FLINK_HOME=${FLINK_HOME:-/opt/soft/flink}
export DATAX_HOME=${DATAX_HOME:-/opt/soft/datax}

export PATH=$HADOOP_HOME/bin:$SPARK_HOME1/bin:$SPARK_HOME2/bin:$PYTHON_HOME/bin:$JAVA_HOME/bin:$HIVE_HOME/bin:$FLINK_HOME/bin:$DATAX_HOME/bin:$PATH

Initialize the Database

DolphinScheduler metadata is stored in the relational database. Currently, supports PostgreSQL and MySQL. If you use MySQL, you need to manually download mysql-connector-java driver (8.0.16) and move it to the lib directory of DolphinScheduler, which is tools/libs/. Let's take MySQL as an example for how to initialize the database:

For mysql 5.6 / 5.7

mysql -uroot -p

mysql> CREATE DATABASE dolphinscheduler DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;

# Replace {user} and {password} with your username and password
mysql> GRANT ALL PRIVILEGES ON dolphinscheduler.* TO '{user}'@'%' IDENTIFIED BY '{password}';
mysql> GRANT ALL PRIVILEGES ON dolphinscheduler.* TO '{user}'@'localhost' IDENTIFIED BY '{password}';

mysql> flush privileges;

For mysql 8:

mysql -uroot -p

mysql> CREATE DATABASE dolphinscheduler DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;

# Replace {user} and {password} with your username and password
mysql> CREATE USER '{user}'@'%' IDENTIFIED BY '{password}';
mysql> GRANT ALL PRIVILEGES ON dolphinscheduler.* TO '{user}'@'%';
mysql> CREATE USER '{user}'@'localhost' IDENTIFIED BY '{password}';
mysql> GRANT ALL PRIVILEGES ON dolphinscheduler.* TO '{user}'@'localhost';
mysql> FLUSH PRIVILEGES;

Then, modify ./bin/env/dolphinscheduler_env.sh to use mysql, change {user} and {password} to what you set in the previous step.

export DATABASE=${DATABASE:-mysql}
export SPRING_PROFILES_ACTIVE=${DATABASE}
export SPRING_DATASOURCE_DRIVER_CLASS_NAME=com.mysql.cj.jdbc.Driver
export SPRING_DATASOURCE_URL="jdbc:mysql://127.0.0.1:3306/dolphinscheduler?useUnicode=true&characterEncoding=UTF-8&useSSL=false"
export SPRING_DATASOURCE_USERNAME={user}
export SPRING_DATASOURCE_PASSWORD={password}

After the above steps done you would create a new database for DolphinScheduler, then run the Shell script to init database:

sh tools/bin/upgrade-schema.sh

Start DolphinScheduler

Use deployment user you created above, running the following command to complete the deployment, and the server log will be stored in the logs folder.

sh ./bin/install.sh

Note: For the first time deployment, there maybe occur five times of sh: bin/dolphinscheduler-daemon.sh: No such file or directory in the terminal, this is non-important information that you can ignore.

Login DolphinScheduler

Access address http://localhost:12345/dolphinscheduler/ui and login DolphinScheduler UI. The default username and password are admin/dolphinscheduler123

Start or Stop Server

# Stop all DolphinScheduler server
sh ./bin/stop-all.sh

# Start all DolphinScheduler server
sh ./bin/start-all.sh

# Start or stop DolphinScheduler Master
sh ./bin/dolphinscheduler-daemon.sh stop master-server
sh ./bin/dolphinscheduler-daemon.sh start master-server

# Start or stop DolphinScheduler Worker
sh ./bin/dolphinscheduler-daemon.sh start worker-server
sh ./bin/dolphinscheduler-daemon.sh stop worker-server

# Start or stop DolphinScheduler Api
sh ./bin/dolphinscheduler-daemon.sh start api-server
sh ./bin/dolphinscheduler-daemon.sh stop api-server

# Start or stop Alert
sh ./bin/dolphinscheduler-daemon.sh start alert-server
sh ./bin/dolphinscheduler-daemon.sh stop alert-server

Note1:: Each server have dolphinscheduler_env.sh file in path <server-name>/conf/dolphinscheduler_env.sh which for micro-services need. It means that you could start all servers by command <server-name>/bin/start.sh with different environment variable from bin/env/dolphinscheduler_env.sh. But it will use file bin/env/dolphinscheduler_env.sh overwrite <server-name>/conf/dolphinscheduler_env.sh if you start server with command /bin/dolphinscheduler-daemon.sh start <server-name>.

Note2:: Please refer to the section of "System Architecture Design" for service usage. Python gateway service is started along with the api-server, and if you do not want to start Python gateway service please disabled it by changing the yaml config python-gateway.enabled : false in api-server's configuration path api-server/conf/application.yaml