I85PREP
⌘ K
Cloudera CDP Generalist study materials
Cloudera CDP Generalist study resources

Cloudera CDP Generalist practice questions

One question at a time. Build your Cloudera CDP Generalist understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Cloudera Data Platform Generalist (domain list not published in source)

In Cloudera Data Platform, which component is responsible for centralized metadata management and governance across data catalogs, including lineage and audit?

  1. Apache Ranger
  2. Apache Knox
  3. Apache Oozie
  4. Apache Atlas Correct answer.

Explanation

Apache Atlas provides the metadata and governance layer in CDP, storing technical and business metadata, tracking lineage, and enabling classification-driven policies. Ranger enforces authorization, Knox handles perimeter security, and Oozie orchestrates workflows, so none of them owns the metadata catalog.

Question 2Cloudera Data Platform Generalist (domain list not published in source)

Which CDP component enforces fine-grained authorization policies such as column masking and row filtering on Hive and HDFS resources?

  1. Apache Ranger Correct answer.
  2. Apache Zeppelin
  3. Apache HBase
  4. Apache Atlas

Explanation

Ranger is the centralized authorization service in CDP; its plugins evaluate policies for HDFS, Hive, HBase, and other services, including masking and row-filter rules. Atlas supplies classifications, but enforcement is performed by the Ranger plugin inside each service.

Question 3Cloudera Data Platform Generalist (domain list not published in source)

What is the primary role of Cloudera Manager in a CDP Private Cloud Base deployment?

  1. Scheduling recurring Oozie workflows
  2. Storing objects in the S3-compatible namespace
  3. Running SQL queries against Hive tables
  4. Provisioning, configuring, and monitoring cluster services Correct answer.

Explanation

Cloudera Manager is the administrative control plane: it installs and configures services, monitors health, manages roles across hosts, and surfaces alerts. Query execution, object storage, and workflow scheduling are handled by other services such as Hive, Ozone, and Oozie.

Question 4Cloudera Data Platform Generalist (domain list not published in source)

In CDP, which storage layer is the native S3-compatible object store designed as a scalable alternative to HDFS?

  1. Apache Kudu
  2. Apache Kafka
  3. Apache Hudi
  4. Apache Ozone Correct answer.

Explanation

Ozone is CDP's distributed object store with an S3-compatible gateway, offering namespace scale beyond HDFS NameNode limits. Kudu is a columnar store for analytics, Hudi is a table format, and Kafka is a streaming platform.

Question 5Cloudera Data Platform Generalist (domain list not published in source)

Which CDP capability lets a data steward apply a classification tag to a table so that Ranger can automatically mask sensitive columns?

  1. YARN queue placement rules
  2. Atlas classification propagated to Ranger policies Correct answer.
  3. Hue saved queries
  4. Knox topology definitions

Explanation

Atlas classifications such as PII can be linked to Ranger tag-based policies, so tagging a column in Atlas causes Ranger to enforce masking without per-column policy authoring. This tag-driven model is the standard governance pattern in CDP.

Question 6Cloudera Data Platform Generalist (domain list not published in source)

A team needs to run a Spark job that reads from and writes to a Kerberized HDFS cluster. Which mechanism allows the job to authenticate without embedding a user password in code?

  1. Plain SASL with no credentials
  2. HTTP basic auth against NameNode
  3. Anonymous LDAP bind
  4. Kerberos keytab with a service principal Correct answer.

Explanation

Non-interactive jobs authenticate to Kerberos using a keytab containing the service principal's encrypted keys, obtained via kinit or the UserGroupInformation login. This avoids storing passwords while still giving the job a valid ticket.

Question 7Cloudera Data Platform Generalist (domain list not published in source)

Which CDP service provides a web-based SQL editor where analysts can browse Atlas metadata, write Hive or Impala queries, and save results?

  1. Apache Flume
  2. Apache Sqoop
  3. Apache Solr
  4. Apache Hue Correct answer.

Explanation

Hue is the browser-based interface bundling a query editor, metadata browser, file browser, and job designer. Solr provides search indexing, Flume ingests streaming data, and Sqoop moved bulk data between Hadoop and relational databases.

Question 8Cloudera Data Platform Generalist (domain list not published in source)

In CDP Data Warehouse, which service is the default SQL engine tuned for low-latency BI queries on large datasets with a shared metadata store alongside Hive?

  1. Apache Storm
  2. Apache Pig
  3. Apache NiFi
  4. Apache Impala Correct answer.

Explanation

Impala is the MPP SQL engine in CDP designed for interactive, low-latency analytic queries, sharing the Hive Metastore for table definitions. Pig is a scripting dataflow tool, NiFi handles data flow automation, and Storm processes streams.

Question 9Cloudera Data Platform Generalist (domain list not published in source)

Which CDP runtime service lets users submit batch applications and manages queue-based resource allocation across a cluster?

  1. Hive Metastore
  2. ZooKeeper
  3. HDFS
  4. YARN Correct answer.

Explanation

YARN is the resource management layer that schedules containers for applications according to configured queues and capacity or fair-share policies. ZooKeeper coordinates distributed services, HDFS stores data, and the Metastore holds table definitions.

Question 10Cloudera Data Platform Generalist (domain list not published in source)

An administrator wants to protect a CDP cluster from unauthorized external access while enabling single sign-on for web UIs. Which component is designed for this perimeter role?

  1. Apache Oozie
  2. Apache Kudu
  3. Apache Knox Correct answer.
  4. Apache Atlas

Explanation

Knox acts as the cluster's edge gateway, terminating TLS, enforcing authentication, and proxying requests to internal UIs and REST APIs. It centralizes SSO and shields internal hosts from direct exposure.

Question 11Cloudera Data Platform Generalist (domain list not published in source)

Which CDP feature allows the same SQL and Spark code to run against on-premises HDFS or cloud object storage without code changes?

  1. Shared data experience with externalized tables and storage connectors Correct answer.
  2. Manual data copies into each engine's local disk
  3. Hard-coded HDFS path rewriting per environment
  4. Separate binaries compiled per storage backend

Explanation

CDP externalizes table metadata and uses storage connectors so the same table definitions and queries reference data on HDFS, Ozone, S3, or ADLS. This decouples compute from storage and enables workload portability across environments.

Question 12Cloudera Data Platform Generalist (domain list not published in source)

A data engineer needs streaming ingestion with exactly-once semantics into HDFS or Kafka topics in CDP. Which service is built for reliable event streaming and topic-based pipelines?

  1. Apache HBase
  2. Apache Kafka Correct answer.
  3. Apache Hive
  4. Apache Tez

Explanation

Kafka is the distributed event streaming platform in CDP, providing durable topics, replication, and consumer-group semantics used for real-time pipelines. Hive, HBase, and Tez are batch query, NoSQL, and execution-engine components respectively.

Question 13Cloudera Data Platform Generalist (domain list not published in source)

Which tool in CDP provides a visual, flow-based interface for automating data movement between systems with provenance tracking?

  1. Apache ZooKeeper
  2. Apache NiFi Correct answer.
  3. Apache Hive
  4. Apache Impala

Explanation

NiFi offers a drag-and-drop canvas of processors for routing, transforming, and tracking data flows, with built-in provenance and back-pressure. The other options are a SQL warehouse layer, a coordination service, and an MPP query engine.

Question 14Cloudera Data Platform Generalist (domain list not published in source)

In CDP, what does the Shared Data Experience (SDX) primarily provide across workloads?

  1. A common JVM heap shared by all services
  2. A unified security and governance context spanning multiple analytic engines Correct answer.
  3. A single global YARN queue for every application
  4. A shared disk volume mounted by all cluster hosts

Explanation

SDX delivers consistent authentication, authorization, audit, and metadata across engines so a policy defined once applies everywhere. It is a logical security and governance layer, not a shared disk, queue, or memory construct.

Question 15Cloudera Data Platform Generalist (domain list not published in source)

Which CDP component stores table schemas and partition metadata consumed by Hive, Impala, and Spark?

  1. Hive Metastore Correct answer.
  2. JournalNode
  3. NodeManager
  4. ResourceManager

Explanation

The Hive Metastore is the shared catalog holding database, table, column, and partition definitions that multiple SQL and processing engines read. ResourceManager and NodeManager belong to YARN, while JournalNodes support HDFS NameNode high availability.

Question 16Cloudera Data Platform Generalist (domain list not published in source)

A security team wants every access to sensitive Hive tables recorded for compliance review. Which CDP capability should they enable?

  1. HDFS block replication increase
  2. Ranger audit logging to a configured destination such as Solr or HDFS Correct answer.
  3. YARN preemption of long-running applications
  4. Atlas lineage-only capture without policy evaluation

Explanation

Ranger plugins emit audit events for each authorization decision, and these can be routed to Solr, HDFS, or a database for review. Lineage alone does not record who accessed what, and replication or preemption settings are unrelated to access auditing.

Question 17Cloudera Data Platform Generalist (domain list not published in source)

Which statement best describes how CDP Data Hub differs from a CDP Data Warehouse environment?

  1. Data Hub provisions customizable clusters for diverse engines, while Data Warehouse offers a managed, tuned SQL experience Correct answer.
  2. Data Hub is on-premises only, while Data Warehouse is cloud only
  3. Data Hub only stores objects, while Data Warehouse only streams events
  4. Data Hub runs without any security integration, while Data Warehouse requires Kerberos

Explanation

Data Hub lets users spin up clusters with a chosen mix of services such as Spark, Kafka, or HBase for varied workloads. Data Warehouse is a curated environment optimized for SQL analytics with Impala and Hive. Both integrate with the same security and governance services.

Question 18Cloudera Data Platform Generalist (domain list not published in source)

An analyst reports that a query against an external table returns no rows although files exist at the location. Which cause is most consistent with CDP external table behavior?

  1. The NameNode is in safe mode cluster-wide
  2. Ranger has disabled the HDFS replication factor
  3. The table's LOCATION points to a different path than where the files reside Correct answer.
  4. The Hive Metastore has run out of YARN containers

Explanation

External tables only read data under their declared LOCATION, so a path mismatch yields empty results even when files exist elsewhere. Metastore scheduling, replication settings, and NameNode safe mode are unrelated to which directory a table scans.

Question 19Cloudera Data Platform Generalist (domain list not published in source)

Which CDP Private Cloud Base capability allows multiple, isolated compute clusters to share a single HDFS or Ozone storage namespace?

  1. Local disk staging of all tables before query
  2. Per-cluster duplicated copies of all datasets
  3. Independent Kerberos realms per workload with no trust
  4. Federated or shared-data architectures with externalized metadata Correct answer.

Explanation

CDP supports separating compute from a shared storage layer, so several clusters can query the same datasets with consistent metadata. Duplicating data, isolating realms without trust, or staging locally all undermine the shared-data model rather than enable it.

Question 20Cloudera Data Platform Generalist (domain list not published in source)

Which configuration ensures that a CDP service's Kerberos principal and keytab are valid for its host before startup?

  1. A shared user password stored in a world-readable file
  2. TLS certificates used in place of Kerberos tickets
  3. A keytab generated for the service principal with matching host FQDN and a valid ticket Correct answer.
  4. Anonymous access enabled on the service endpoint

Explanation

Kerberized services need a keytab containing the correct principal, whose host component matches the machine's fully qualified domain name, plus a valid ticket for authentication. Passwords in readable files, anonymous access, and TLS certificates do not substitute for Kerberos identity.

Question 21Cloudera Data Platform Generalist (domain list not published in source)

Which CDP component provides a columnar NoSQL store optimized for fast random reads and writes, often used for real-time dashboards?

  1. Apache Ozone
  2. Apache Kudu Correct answer.
  3. Apache Atlas
  4. Apache Knox

Explanation

Kudu stores data in columns with primary keys, supporting low-latency random access and updates alongside analytical scans. Ozone is object storage, Atlas is metadata governance, and Knox is an edge gateway.

Question 22Cloudera Data Platform Generalist (domain list not published in source)

What is the purpose of a Ranger tag-based policy in CDP governance?

  1. To schedule recurring Oozie coordinator jobs
  2. To encrypt data at rest using HDFS transparent encryption
  3. To replicate HDFS blocks across racks
  4. To enforce access rules based on Atlas classifications rather than explicit resource paths Correct answer.

Explanation

Tag-based policies bind authorization to classifications like PII that Atlas assigns to resources, so protection follows the data as it moves between tables. Encryption, workflow scheduling, and block replication are separate mechanisms.

Question 23Cloudera Data Platform Generalist (domain list not published in source)

Which service in CDP is responsible for coordinating distributed components such as leader election and configuration consistency?

  1. Apache ZooKeeper Correct answer.
  2. Apache Hive
  3. Apache NiFi
  4. Apache Hue

Explanation

ZooKeeper provides a consistent, highly available coordination service used by components for leader election, configuration management, and distributed locks. Hive, NiFi, and Hue are query, dataflow, and UI services respectively.

Question 24Cloudera Data Platform Generalist (domain list not published in source)

A cluster must keep serving reads and writes when the active NameNode host fails. Which HDFS mechanism enables this?

  1. NameNode high availability with JournalNodes and automatic failover Correct answer.
  2. Disabling permissions checks on the NameNode
  3. Running the Secondary NameNode as a hot standby for writes
  4. Increasing the replication factor of the NameNode metadata directory

Explanation

NameNode HA uses a standby NameNode synchronized through JournalNodes, with ZooKeeper-based automatic failover so clients reconnect to the new active node. The Secondary NameNode only checkpoints edits and cannot take over writes.

Question 25Cloudera Data Platform Generalist (domain list not published in source)

Which CDP capability supports querying data across both on-premises HDFS tables and cloud object storage tables from a single engine?

  1. Copying all cloud data into HDFS nightly
  2. Running separate engines with no shared catalog
  3. External tables registered in the shared Hive Metastore Correct answer.
  4. Using only local filesystem paths in queries

Explanation

Because external table definitions in the shared Metastore can point to HDFS, Ozone, S3, or ADLS locations, one engine can query data wherever it resides. Nightly copies, isolated catalogs, and local-only paths all prevent unified access.

Question 26Cloudera Data Platform Generalist (domain list not published in source)

In CDP, which component is the centralized web UI where administrators create, publish, and audit Apache Ranger policies for Hive, HDFS, and Kafka?

  1. Ranger Admin Correct answer.
  2. YARN ResourceManager
  3. Atlas
  4. Knox Gateway

Explanation

Ranger Admin is the policy management server in Ranger; administrators define, version, and audit access policies there, and the plugins embedded in each service enforce them. Atlas is a metadata and lineage catalog, not a policy authoring console.

Question 27Cloudera Data Platform Generalist (domain list not published in source)

Which CDP capability lets an administrator define a virtual cluster boundary that groups hosts, services, and configurations so multiple workloads can be provisioned and torn down independently on shared infrastructure?

  1. Replication Manager policies
  2. Clusters (Virtual Cluster) in Cloudera Manager Correct answer.
  3. HDFS erasure coding policies
  4. YARN Node Labels

Explanation

In CDP Private Cloud Base, the Clusters feature in Cloudera Manager creates virtual clusters: isolated sets of services and roles that can be created, scaled, and deleted without affecting other virtual clusters on the same hosts. This is the mechanism behind multi-tenant compute sharing one storage layer.

Question 28Cloudera Data Platform Generalist (domain list not published in source)

A CDP environment must let external BI tools authenticate users through the customer's corporate identity provider using SAML, while internal service-to-service calls continue to use Kerberos. Which statement describes the correct design?

  1. Kerberos must be disabled before SAML can be enabled in a CDP environment
  2. SAML is only supported for the Cloudera Manager admin console and cannot be used by data services
  3. SAML federation handles interactive web SSO, and Kerberos remains for non-interactive service authentication Correct answer.
  4. SAML replaces Kerberos entirely, so all daemons must be reconfigured to use SAML tokens for RPC

Explanation

CDP supports SAML-based single sign-on for browser and JDBC/ODBC style interactive access while Kerberos continues to secure daemon-to-daemon and service RPC authentication. The two mechanisms coexist rather than substituting for one another.

Question 29Cloudera Data Platform Generalist (domain list not published in source)

Which CDP service provides a REST API and UI for defining, scheduling, and monitoring data replication and migration jobs between HDFS, Hive, and cloud storage?

  1. YARN Capacity Scheduler
  2. ZooKeeper
  3. Hue
  4. Replication Manager Correct answer.

Explanation

Replication Manager orchestrates HDFS and Hive replication, including metadata such as table definitions, and exposes job status through a UI and API. It is the standard CDP tool for disaster recovery and migration copy jobs.

Question 30Cloudera Data Platform Generalist (domain list not published in source)

An operator wants to add a new DataNode to an existing CDP Private Cloud Base cluster with minimal manual steps. Which Cloudera Manager workflow supports this?

  1. Deleting and recreating the HDFS service with the larger host list
  2. Manually editing hdfs-site.xml on every node and restarting the NameNode
  3. Running the Replication Manager migration job against the new host
  4. Add Hosts wizard followed by assigning the HDFS DataNode role to the new host Correct answer.

Explanation

The Add Hosts wizard installs the Cloudera Manager agent and parcels on new hosts, and then roles such as DataNode are assigned through the cluster's service configuration. This is the supported, repeatable way to scale storage capacity.

Question 31Cloudera Data Platform Generalist (domain list not published in source)

Which CDP component presents a unified SQL endpoint and connection URL that routes JDBC and ODBC clients to the appropriate Hive or Impala engine?

  1. HiveServer2 (or the Hive/Impala JDBC endpoint exposed via Cloudera Data Warehouse) Correct answer.
  2. The NameNode RPC port
  3. The Ranger Admin audit socket
  4. The ResourceManager scheduler address

Explanation

SQL clients connect to a HiveServer2 or Impala endpoint, which accepts JDBC/ODBC connections and executes queries against the underlying engine and metastore. The NameNode, ResourceManager, and Ranger ports serve entirely different protocols.

Question 32Cloudera Data Platform Generalist (domain list not published in source)

A data engineer needs to run an Apache Flink or custom YARN application alongside existing MapReduce and Spark workloads while preventing one workload from starving the others. Which CDP mechanism should be configured?

  1. Ranger row-filter policies
  2. HDFS block placement policies
  3. Atlas lineage tags
  4. YARN queues with Capacity Scheduler or Fair Scheduler weights and limits Correct answer.

Explanation

YARN queues, configured through the Capacity or Fair Scheduler, divide cluster resources among workload classes and enforce minimum guarantees and maximum caps. This prevents a single application type from consuming all containers.

Question 33Cloudera Data Platform Generalist (domain list not published in source)

Which statement accurately describes the role of Apache Ozone in CDP?

  1. A columnar analytics database that replaces Impala
  2. A distributed object store with an S3-compatible API that can serve as the storage layer for CDP workloads Correct answer.
  3. A policy engine that replaces Ranger
  4. A message broker used for streaming ingestion

Explanation

Ozone is a scalable object store that provides S3-compatible access and namespace management, and it can back HDFS-compatible workloads in CDP. It is a storage service, not a query engine, broker, or authorization system.

Question 34Cloudera Data Platform Generalist (domain list not published in source)

An administrator must ensure that when a user queries a Hive table, only rows for the user's own region are returned, regardless of the query written. Which Ranger capability implements this?

  1. Row-level filter policies Correct answer.
  2. HDFS encryption zones
  3. Column masking policies
  4. Atlas classification propagation

Explanation

Ranger row filter policies attach a predicate to a table for specified users or groups, and the engine appends that predicate so only matching rows are visible. Column masking hides values within a column but does not restrict which rows appear.

Question 35Cloudera Data Platform Generalist (domain list not published in source)

What is the primary function of the Hive Metastore in a CDP deployment?

  1. Storing the actual row data for managed tables on local disk
  2. Managing Kerberos keytabs for cluster hosts
  3. Storing table, column, partition, and location definitions that engines use to resolve datasets Correct answer.
  4. Scheduling YARN containers for Hive queries

Explanation

The Hive Metastore is a relational catalog of database, table, column, partition, and storage-location metadata, consumed by engines such as Hive, Impala, and Spark. Row data lives in HDFS or object storage, not in the metastore.

Question 36Cloudera Data Platform Generalist (domain list not published in source)

A team wants to use a single SQL statement to join a Hive table stored in HDFS with a table whose data resides in an S3 bucket. Which CDP capability makes this possible?

  1. The shared metastore cataloging both tables, with engines reading each table's configured storage location Correct answer.
  2. Converting the S3 table into a local HDFS file before querying
  3. Disabling the metastore so engines can read raw files directly
  4. Requiring both tables to be copied into the same HDFS directory first

Explanation

CDP engines resolve each table through the shared metastore and read from the location recorded for that table, whether HDFS or object storage. A single query can therefore join tables across storage backends without a copy step.

Question 37Cloudera Data Platform Generalist (domain list not published in source)

Which CDP service is designed to ingest streaming event data at scale and durably buffer it in topics for downstream consumers?

  1. Apache Sentry
  2. Apache Kafka Correct answer.
  3. Apache Atlas
  4. Apache Knox

Explanation

Kafka is a distributed publish-subscribe log that stores records in partitioned topics so producers and consumers can operate independently and at high throughput. It is the standard streaming backbone in CDP.

Question 38Cloudera Data Platform Generalist (domain list not published in source)

An analyst needs to explore data interactively with SQL and also browse table metadata, lineage, and saved queries from one web interface. Which CDP component best fits?

  1. JournalNode
  2. NodeManager
  3. ZooKeeper
  4. Hue Correct answer.

Explanation

Hue is a browser-based workbench that offers SQL editors for Hive and Impala, metadata and lineage browsing backed by Atlas, and saved query management. The other components are infrastructure daemons with no end-user analytics UI.

Question 39Cloudera Data Platform Generalist (domain list not published in source)

Which mechanism does HDFS use to keep a standby NameNode ready to take over automatically when the active NameNode fails?

  1. Secondary NameNode checkpointing every hour
  2. DataNode block scanner reports
  3. NameNode high availability with ZooKeeper-based automatic failover and JournalNode edit-log sharing Correct answer.
  4. HDFS trash checkpoint interval

Explanation

In HDFS HA, the active and standby NameNodes share edit logs through JournalNodes, and ZooKeeper with the ZKFC controller detects failure and triggers automatic failover. The Secondary NameNode only performs checkpointing and cannot serve as a failover target.

Question 40Cloudera Data Platform Generalist (domain list not published in source)

A cluster has Kerberos enabled, and a new Hue service must authenticate to HiveServer2. Which artifact must be provisioned for the Hue principal before the service can start successfully?

  1. A Ranger audit log file for the Hue host
  2. A PEM certificate signed by the cluster's CA for each Hive table
  3. A keytab containing the Hue principal's key, readable only by the Hue process user Correct answer.
  4. An Atlas classification named after the Hue principal

Explanation

Kerberized services authenticate using a principal and a keytab that holds the principal's encrypted keys; the keytab must be generated by the KDC tooling and secured with restrictive permissions. Without a valid keytab, the service cannot obtain a ticket and startup fails.

Question 41Cloudera Data Platform Generalist (domain list not published in source)

Which CDP component maintains a searchable catalog of datasets, tracks column-level lineage, and lets stewards attach business classifications to assets?

  1. Apache Oozie
  2. Apache Atlas Correct answer.
  3. Apache Ranger
  4. Apache Tez

Explanation

Atlas is the metadata and governance catalog in CDP; it ingests technical metadata, computes lineage, and supports classifications that can drive tag-based policies. Ranger enforces access, while Oozie and Tez are workflow and execution engines.

Question 42Cloudera Data Platform Generalist (domain list not published in source)

A steward wants to classify all tables containing personally identifiable information and have that classification automatically drive access restrictions without rewriting file-path policies. Which combination enables this?

  1. YARN queue ACLs per table
  2. Impala query hints per user
  3. Atlas classifications synced to Ranger, with tag-based Ranger policies Correct answer.
  4. HDFS encryption zones with a new key per table

Explanation

Atlas classifications can be synchronized to Ranger, which evaluates tag-based policies against those classifications. This lets stewards classify assets once and have enforcement follow the tag rather than a hard-coded path or database name.

Question 43Cloudera Data Platform Generalist (domain list not published in source)

Which statement best explains why CDP separates compute clusters from the shared storage and catalog layer?

  1. Multiple compute clusters can be created and destroyed independently while data and metadata persist in shared storage Correct answer.
  2. Storage must be deleted whenever a compute cluster is decommissioned
  3. Each compute cluster must own a private copy of all data to function
  4. The catalog is rebuilt from scratch each time a cluster starts

Explanation

Decoupling compute from storage lets teams spin up isolated compute environments for different workloads or teams while the data and metastore remain in a durable shared layer. This reduces duplication and shortens environment provisioning.

Question 44Cloudera Data Platform Generalist (domain list not published in source)

Which CDP feature provides a centralized audit repository where Ranger records allowed and denied access decisions for later compliance review?

  1. The YARN application timeline store
  2. The Hive query compilation log
  3. Ranger audit with a configured audit destination such as HDFS or a database Correct answer.
  4. The NameNode safe mode log

Explanation

Ranger plugins emit audit events for each authorization decision, and Ranger Admin stores them in a configured destination such as Solr, HDFS, or a relational database for search and retention. This provides the evidence trail auditors require.

Question 45Cloudera Data Platform Generalist (domain list not published in source)

An administrator must give a group read access to one HDFS directory while denying access to its parent directory's other children. Which approach is most appropriate in a secure CDP cluster?

  1. A Ranger HDFS policy granting the group read on that path, with parent paths traversable but not listable Correct answer.
  2. Disabling HDFS permissions checks cluster-wide
  3. Granting the group superuser status so path permissions no longer apply
  4. Setting the directory to world-readable recursively

Explanation

Ranger HDFS policies can grant access to a specific path while execute-only traversal on ancestors allows the group to reach it without listing siblings. This follows least privilege and avoids broad exposure.

Question 46Cloudera Data Platform Generalist (domain list not published in source)

Which CDP component acts as the single entry point that proxies authenticated HTTP access to service UIs and REST APIs from outside the cluster?

  1. Apache Knox Gateway Correct answer.
  2. Apache Atlas
  3. Apache ZooKeeper
  4. HDFS Balancer

Explanation

Knox sits at the cluster edge, authenticates external users through federation or LDAP, and proxies requests to internal service endpoints without exposing each UI directly. This reduces the external attack surface.

Question 47Cloudera Data Platform Generalist (domain list not published in source)

A workload requires low-latency interactive SQL over large Parquet tables with many concurrent BI users. Which CDP engine characteristic makes Impala suitable?

  1. An engine that only supports batch MapReduce jobs
  2. A single-node engine that processes all queries serially
  3. A massively parallel execution engine with long-running daemons that avoid per-query startup overhead Correct answer.
  4. An engine that stores data exclusively in local memory with no disk reads

Explanation

Impala runs as a distributed set of long-lived daemons that execute queries in parallel across nodes, avoiding the job startup cost of batch frameworks. That design yields the sub-second to seconds latency BI users expect.

Question 48Cloudera Data Platform Generalist (domain list not published in source)

Which statement describes how CDP handles a query against an external table whose underlying files were deleted outside the metastore?

  1. Ranger recreates the files as part of policy enforcement
  2. The engine silently substitutes a backup table with the same name
  3. The query fails or returns no rows because the metastore location no longer resolves to existing files Correct answer.
  4. The metastore automatically regenerates the missing files from its own copy

Explanation

An external table's metadata records a location but does not own the data, so deleting those files leaves the table definition intact while queries find nothing to read. Recovery requires restoring the files or repairing metadata, not a policy action.

Question 49Cloudera Data Platform Generalist (domain list not published in source)

Which CDP capability allows a Spark job running in one compute cluster to read a Hive table whose metadata is registered in the shared metastore?

  1. Granting the Spark user HDFS superuser rights
  2. Disabling the metastore and reading raw files with no schema
  3. Configuring the Spark session to use the shared Hive Metastore URI so table names resolve to catalog entries Correct answer.
  4. Copying the table's CREATE statement into the Spark code as a hard-coded schema

Explanation

When Spark is pointed at the shared metastore, table names resolve through the catalog to their schemas and storage locations, so the job can read the table without redefining it. Superuser rights or hard-coded schemas are not required for this.

Question 50Cloudera Data Platform Generalist (domain list not published in source)

An organization wants to prove that access to a regulated table was denied to an unauthorized group last quarter. Which evidence should they retrieve?

  1. The YARN application logs for the quarter
  2. The HDFS fsck report for the table's directory
  3. The Atlas lineage graph for the table
  4. Ranger audit records showing the denied authorization decision for that table and group Correct answer.

Explanation

Ranger audit records capture each authorization decision, including denials, with user, resource, and timestamp, which is exactly the evidence needed. Fsck, application logs, and lineage describe data health, execution, and provenance rather than access decisions.