CDP (Big Data Hadoop Administrator) 2 — Questions and Answers
Question 1: Which HDFS configuration parameter controls the minimum percentage of DataNodes that must successfully write a block before the client receives acknowledgment?
- dfs.replication.min (Correct answer)
- dfs.datanode.failed.volumes.tolerated
- dfs.namenode.replication.min
- dfs.blocksize
Correct answer: dfs.replication.min
dfs.replication.min sets the minimum number of replicas that must be written before the write is considered successful.
Question 2: In YARN, what is the role of the ApplicationMaster?
- Managing all cluster resources globally
- Negotiating resources and tracking task status for a specific application (Correct answer)
- Scheduling jobs across the cluster
- Monitoring DataNode health
Correct answer: Negotiating resources and tracking task status for a specific application
The ApplicationMaster is a per-application process that negotiates resources from the ResourceManager and works with NodeManagers to execute tasks.
Question 3: Which command correctly increases the replication factor of an existing HDFS file to 5?
- hdfs dfs -setrep -w 5 /path/to/file (Correct answer)
- hdfs dfs -rep 5 /path/to/file
- hdfs fsck -replication 5 /path/to/file
- hdfs namenode -setReplication 5 /path/to/file
Correct answer: hdfs dfs -setrep -w 5 /path/to/file
The hdfs dfs -setrep command changes the replication factor; -w waits for the replication to complete.
Question 4: A Hadoop cluster has frequent 'GC overhead limit exceeded' errors on the NameNode. What is the most appropriate first corrective action?
- Increase dfs.namenode.handler.count
- Increase heap memory allocated to the NameNode via HADOOP_NAMENODE_OPTS (Correct answer)
- Reduce the HDFS block size
- Add more DataNodes to the cluster
Correct answer: Increase heap memory allocated to the NameNode via HADOOP_NAMENODE_OPTS
GC overhead errors on the NameNode indicate insufficient heap memory; increasing it via HADOOP_NAMENODE_OPTS is the direct fix.
Question 5: What does the 'safe mode' state in HDFS prevent?
- New user logins to the cluster
- Any modifications to the file system namespace (Correct answer)
- DataNodes from registering with the NameNode
- Execution of MapReduce jobs
Correct answer: Any modifications to the file system namespace
In safe mode the NameNode is read-only; it does not allow any modifications to the HDFS namespace.
Question 6: Which Hadoop component is responsible for storing job history and providing a REST API to query completed MapReduce jobs?
- ResourceManager
- JobHistoryServer (Correct answer)
- ApplicationMaster
- NodeManager
Correct answer: JobHistoryServer
The JobHistoryServer retains logs and metrics for completed MapReduce jobs and exposes them via a web UI and REST API.
Question 7: In a Kerberos-secured Hadoop cluster, what artifact does a service use to authenticate without a human entering a password?
- SSL certificate
- Keytab file (Correct answer)
- LDAP token
- Knox gateway token
Correct answer: Keytab file
A keytab file contains encrypted Kerberos principal credentials, allowing services to obtain Kerberos tickets automatically.
CDP (Big Data Hadoop Administrator) 2
Which HDFS configuration parameter controls the minimum percentage of DataNodes that must successfully write a block before the client receives acknowledgment?