Start and stop the Arenadata Hadoop cluster
Follow the instructions below to learn how to start and stop an Arenadata Hadoop (ADH) cluster.
To manage your ADH cluster, you need to have Arenadata Cluster Manager (ADCM) and ADH installed. Follow Get started with Arenadata Hadoop for additional information.
Starting the cluster
Starting the monitoring cluster (optional)
If you have installed a monitoring cluster and want to use it, first start it (before you start the main cluster) as follows:
-
On the CLUSTERS tab, find your monitoring cluster (for example, called Monitoring as in the following screenshot) and click the
icon in the Action column. In the drop-down menu, choose Start.
Starting the monitoring cluster -
Click Run to confirm. You can track the start of a monitoring cluster using the task status in the upper right corner of the screen.
Confirmation of the start request
Starting the main cluster
Start the ADH cluster using one of the following options:
-
Start the entire cluster so that it automatically starts all the services installed on it in the correct order. This operation is similar to the one you use to start the monitoring cluster.
-
Start only the required cluster services one by one in the correct order.
To keep the cluster startup process under your control, start its services step by step, following this instruction:
-
On the CLUSTERS tab, click your ADH cluster to open its profile.
-
In the left-hand side navigation menu, select Services.
Cluster services -
Start the services manually in the right order:
-
Monitoring [1] (if you use the monitoring cluster)
-
Zookeeper [2]
-
MySQL [3]
-
HDFS [4]
-
YARN [5]
-
HBase [6]
-
Hive [7]
-
Spark [8]
-
Flink [9]
-
Solr [10]
-
Sqoop [11]
-
Airflow [12]
Services in the cluster are split into optional and required. The services required by the most other Hadoop services are:-
Zookeeper
-
HDFS
-
YARN
Start only one service at a time. To start a selected service, click the
icon and choose Start for the selected service.
-
-
Click Run to confirm the start of the service.
Confirmation of the start request -
Wait until all required Hadoop services are started.
Start optional services if needed.
You can track the launch status of services via the special symbol in the Status column or via task status in the upper right corner of the screen.
Stopping the cluster
As with startup, you can either stop the entire cluster with all its services automatically, or stop each service manually in the correct sequence.
To stop the entire cluster, find this cluster on the CLUSTERS tab, click the icon in the Actions column, and select Stop in the drop-down menu.
To keep this process under your control, stop its services step by step, following this instruction:
NOTE
This operation is similar to the operation of starting the services, yet in reverse order.
|
-
Stop the services in the following order:
-
Airflow [1]
-
Sqoop [2]
-
Solr [3]
-
Flink [4]
-
Spark [5]
-
Hive [6]
-
HBase [7]
-
YARN [8]
-
HDFS [9]
-
MySQL [10]
-
Zookeeper [11]
-
Monitoring [12]
On the CLUSTERS tab, select your cluster, then navigate to Services, and find the service. In the Actions column, click theicon and choose the Stop action.
-
-
Click Run to confirm the selected command.
Confirmation of the stop request