+++ title = “Deploy Guide” weight = 1 chapter = true +++

Scheduler deployment steps

1. Start ElasticJob-Cloud-Scheduler and Mesos, and specify ZooKeeper as the registry.
2. Start Mesos Master and Mesos Agent.
3. Unzip elasticjob-cloud-scheduler-${version}.tar.gz.
4. Run bin\start.sh to start ElasticJob-Cloud-Scheduler.

Job deployment steps

1. Ensure that ZooKeeper, Mesos Master/Agent and ElasticJob-Cloud-Scheduler have been started correctly.
2. Place the tar.gz file of the packaging job in a network accessible location, such as ftp or http. The main method in the packaged tar.gz file needs to call the JobBootstrap.execute method provided by ElasticJob-Cloud.
3. Use curl command to call RESTful API to publish applications and register jobs. For details: Configuration

Scheduler configuration steps

Modify the conf\elasticjob-cloud-scheduler.properties to change the system configuration.

Configuration description:

Attribute Name	Required	Default	Description
hostname	yes		The real IP or hostname of the server, cannot be 127.0.0.1 or localhost
user	no		User name used by Mesos framework
mesos_url	yes	zk://127.0.0.1:2181/mesos	Zookeeper url used by Mesos
zk_servers	yes	127.0.0.1:2181	Zookeeper address used by ElasticJob-Cloud
zk_namespace	no	elasticjob-cloud	Zookeeper namespace used by ElasticJob-Cloud
zk_digest	no		Zookeeper digest used by ElasticJob-Cloud
http_port	yes	8899	Port used by RESTful API
job_state_queue_size	yes	10000	The maximum value of the accumulation job, the accumulation job exceeding this threshold will be discarded. Too large value may cause ZooKeeper to become unresponsive, and should be adjusted according to the actual measurement
event_trace_rdb_driver	no		Driver of Job event tracking database
event_trace_rdb_url	no		Url of Job event tracking database
event_trace_rdb_username	no		Username of Job event tracking database
event_trace_rdb_password	no		Password of Job event tracking database
auth_username	no	root	API authentication username
auth_password	no	pwd	API authentication password

---

• Stop: No stop script is provided, you can directly use the kill command to terminate the process.