Unit 2: Hadoop Architecture - Practice Quiz

1 Which type of architecture is traditionally used by Hadoop?

2 Which two components form the core of classic Hadoop architecture?

3 What does HDFS stand for?

4 How does HDFS normally store a large file?

5 What is the main purpose of block replication in HDFS?

6 Which two main phases are used in the MapReduce programming model?

7 What is the primary role of the Map phase?

8 What is the primary role of the Reduce phase?

9 In MapReduce, data is commonly processed in which form?

10 What is the process of transferring mapper output to reducers called?

11 What does a mapper produce before the Reduce phase begins?

12 What is the main responsibility of the HDFS NameNode?

13 Which information is maintained by the NameNode?

14 What is the main responsibility of an HDFS DataNode?

15 What does a DataNode regularly send to the NameNode to show that it is active?

16 In classic Hadoop MapReduce, what is the main role of the JobTracker?

17 Which component submits progress information to the JobTracker in classic Hadoop?

18 What does a TaskTracker do in classic Hadoop MapReduce?

19 Which command pattern correctly runs a Hadoop word-count JAR?

20 Which command can display the first reducer output file from an HDFS word-count result directory named output?

21 A Hadoop cluster must process a large dataset while minimizing network traffic. Which architectural decision best supports this goal?

22 A client submits a MapReduce job that reads a file from HDFS. Which sequence best describes the main interaction?

23 An HDFS file is divided into 5 blocks and stored with a replication factor of 3. Assuming every block is fully replicated, how many block replicas are stored in total?

24 A 300 MB file is stored in HDFS using a block size of 128 MB. How many HDFS blocks are required?

25 A DataNode containing one replica of an HDFS block becomes unavailable. Other replicas remain accessible. What is the expected immediate result?

26 A MapReduce program calculates total sales for each product. The mapper emits (product, saleAmount). What should the reducer do?

27 A job has many repeated mapper keys and performs an associative sum. Which feature can reduce the amount of data transferred during shuffle?

28 A MapReduce job uses four reducers. What determines which reducer receives a particular intermediate key?

29 In a word-count job, a mapper emits (cloud, 1) three times and (data, 1) twice. After shuffle and sort, what input is logically presented to reducers?

30 Which MapReduce component defines how input data is divided into logical splits and converted into key-value records?

31 A mapper task fails after producing intermediate output, but before reducers finish fetching it. What normally happens?

32 A user renames an HDFS file without changing its contents. Which node primarily updates the information for this operation?

33 The NameNode stops receiving heartbeats from a DataNode for an extended period. What action is most appropriate?

34 During an HDFS read, a client discovers that the nearest replica is unavailable. What should the client normally do?

35 Which message allows a DataNode to inform the NameNode about the HDFS blocks it currently stores?

36 In Hadoop MapReduce version 1, a TaskTracker repeatedly fails tasks assigned to it. Which JobTracker response best maintains job progress?

37 The JobTracker has two available TaskTrackers, but only one stores a local replica of the next map task's input block. Which placement is generally preferred?

38 In MapReduce version 1, how does a TaskTracker normally obtain new work from the JobTracker?

39 A word-count job has written results to /output/wc in HDFS. Which command displays all reducer output part files on the terminal?

40 A word-count command fails because its intended HDFS output directory already exists. Which action usually resolves the issue before rerunning the job?

41 In a classic Hadoop 1.x cluster, a worker remains reachable and continues sending TaskTracker heartbeats, but its DataNode process has stopped. Which system behavior is most accurate?

42 A 300 MiB HDFS file uses a 128 MiB block size and replication factor 3. Ignoring checksums and metadata, which statement is correct?

43 An HDFS file has replication factor 3, and all DataNodes are healthy. The NameNode reports that all three replicas of one block are on the same rack. What is the most important consequence?

44 A MapReduce job reads 10,000 files of 1 KiB each from HDFS. The total payload is small, yet startup and metadata overhead are severe. Which explanation best identifies the structural cause?

45 Which statement correctly distinguishes a SecondaryNameNode checkpoint from NameNode failover in classic Hadoop 1.x?

46 After a NameNode restart, HDFS enters safe mode because the configured threshold of reported blocks has not yet been reached. Which action is expected during this period?

47 A client detects a checksum mismatch while reading one replica of an HDFS block, while another replica is healthy. What is the expected recovery path?

48 An administrator wants to retire a healthy DataNode without reducing the durable replication of its blocks. Which procedure has the intended semantics?

49 A large text file has HDFS block boundaries that fall in the middle of several lines. With a standard line-oriented TextInputFormat, why are records normally neither lost nor duplicated?

50 A job uses four reducers and the default conceptual partition rule . Keys A and B have different hash codes but compare equal under the reducer's grouping comparator. What must be true for correctness?

51 A mapper emits (department, (salary, 1)). A proposed combiner replaces all salaries for each department with (averageSalary, 1), and the reducer averages its received salary values. Why can this produce an incorrect global average?

52 A reducer writes directly to an external payment service, and speculative execution is enabled. Two attempts of the same reduce task both reach the service before one attempt is discarded. Which design property is required to prevent duplicate payments?

53 A secondary-sort job uses composite keys (customer, timestamp) and must deliver each customer's records to one reducer in ascending timestamp order. Which configuration is correct?

54 A job has 12 map tasks and 5 reduce tasks. Every mapper emits data, but the partitioner assigns no keys to one reducer. How many final reducer part files are normally created after successful completion?

55 In a default classic Hadoop 1.x deployment without configured recovery or external high availability, what is the principal effect of JobTracker failure?

56 For a map task whose input block has replicas on nodes N1, N2, and N3, no map slot is immediately free on those nodes, but a slot is free on another node in the same rack. Which scheduling choice best reflects delay scheduling for locality?

57 A classic MRv1 TaskTracker is configured with 8 map slots and 2 reduce slots. All map slots are busy, both reduce slots are idle, and another map task is waiting. Under static slot configuration, what normally happens?

58 A TaskTracker stops heartbeating after launching several attempts. The attempts may still be running at the operating-system level. What does the JobTracker normally do after the worker timeout?

59 The command hadoop jar hadoop-mapreduce-examples.jar wordcount /books /wc-out fails immediately with a message that /wc-out already exists. Which remedy preserves the existing output while allowing a new run?

60 A WordCount job completes with three reducers. Running hadoop fs -cat /wc-out/part-r-00000 shows only some words, even though the input contained many more. What is the correct way to inspect the complete reducer output from the command line?