Your running cluster will be having only core services (HDFS, YARN, Zookeeper) or handful of services and your task is to add a specific service to the cluster.
To add a service:
Go to CM – click the drop down box near the cluster – select Add service.
You will get a list of services available to add depends on the CDH version. Select the service and click continue.
Some services are dependent on other service and they can’t be added without adding the dependency service first.
Impala has a dependency of Hive. You can’t add impala service without adding hive first.
Hue has a dependency of HDFS, MR, Hive.
Kafka service requires Kafka parcel has to be installed and activated. If not installed, it will fail while adding the service. So if you’re asked to add Kafka service, please check whether Kafka parcel is installed and activated, then proceed to add service.
After selecting continue, you’ll be asked to customize role assignments for the service.
Hive: Hive Metastore server, hiveserver2, gateway
Impala: Impala daemon, Statestore, Catalog server
HBase: Master, Regionserver, Thrift server
Spark: History server, gateway
HDFS: Datanode, gateway
YARN: Nodemanager, Jobhistory server, gateway
Most likely this information will be given in the problem description, i.e on which host a particular role of the service has to be installed. If some role information is not given then you needn’t to assign that role. Then select continue, verify the service configurations (log dirs path, resources(memory) adjustment etc) and click continue, the service will be added to the cluster.
This is a pretty straightforward process, but you may encounter challenges after adding the service. Like it may show some warnings/errors, one of the roles is not starting etc., You have to identify the problem and rectify it.
Once the service is added, ensure that service is up and running, and no warnings/errors showing up in the service status. Please note that the answer will be marked wrong, if the service is not running.
Since it’s an administrator exam, you won’t be asked to run any Hive/Impala queries. You needn’t to run or test any jobs to check the service unless you’re asked to.
Thus we covered how to Add a service using Cloudera Manager.
Use the comments section below to post your doubts, questions and feedback.
Please follow my blog to get notified of more certification related posts, exam tips, etc.
Hi Kannan, Can you please share procedure to add kafka &impala and configure them?. I would really appreciate.