Ans: As I said earlier Hadoop is open source software where Java framework is used to store, transfer, and clustering the data is called big data. This type of software technology offers huge storage management for any kind of data. Hadoop big data helps in processing enormous data power and offers a mechanism to handle limitless tasks or operations.
Ans: The below table will explain the major differences between Hadoop and Spark;
|Hadoop offers HDFS dedicated storage||With Spark none of the dedicated storage is possible|
|Hadoop offers an Average speed of processing||Spark provides an excellent speed of processing|
|Hadoop supports tools and libraries||Spark supports tools like core, SQL, MLlib, and Graphx|
Ans: The following are the important features of Hadoop;
Ans: The following points will explain how you can differentiate Hadoop from parallel computing system;
Ans: There are 3 types of modes available in Hadoop; such as
Ans: Hadoop’s distributed cache is a type of service provided by MapReduce Framework used to manage cache file systems.
This distributed cache reduces the execution time and it’s also possible to access the data.
Here you can store the data in arrays and hash map.
Ans: There are three types of input formats available in Hadoop;
Ans: The three core methods of Hadoop reducer such as;
Public void setup (context_name)
Public void reduce (key, value, context)
Public void cleanup (context_name)
Ans: Sequence files are extensively used to reduce the input/output formats. This is a flat-file that contains binary key-value pairs. Usually, map outputs are stored in the Sequence file. This sequence file contains three classes such as reader class, writer class, and sorter class.
Ans: Job tracker is a resource management system used to manage the task tracker. This is also used to track the task’s progress and fault tolerance. The job tracker is also used to communicate with the Name node mainly used to identify the location of data.
This can be done by using the following command such as
Ps –ef I grep –I ResourceManager
Ps –ef I grep –iNodemanager
Ans: To achieve this compression method,
Conf. set (“mapreduce.map.output.compress” , true)
Conf. set (“mapreduce. Output.fileoutputmat.compress”, false)
Ans: By writing the query:
Hive > insert overwrite directory ‘/’ select * from emp;
Ans: Below are the three different modes of Hadoop such as;
Ans: It works on MapReduce, and it is devised by Google.
Ans: Map reduce is an algorithm or concept to process Huge amount of data in a faster way. As per its name you can divide it Map and Reduce.
The main MapReduce job usually splits the input data-set into independent chunks. (Big data sets in the multiple small datasets)
MapTask: will process these chunks in a completely parallel manner (One node can process one or more chunks).The framework sorts the outputs of the maps.
Reduce Task : And the above output will be the input for the reducetasks, produces the final result.
Your business logic would be written in the MappedTask and ReducedTask. Typically both the input and the output of the job are stored in a file-system (Not database). The framework takes care of scheduling tasks, monitoring them and re-executes the failed tasks
Ans: The MapReduce framework consists of a single master JobTracker and multiple slaves, each cluster-node will have one TaskTracker. The master is responsible for scheduling the jobs' component tasks on the slaves, monitoring them and re-executing the failed tasks. The slaves execute the tasks as directed by the master.
Ans: Minimally an Hadoop application would have following components.
Input location of data
Output location of processed data.
A map task.
A reduced task.
The Hadoop job client then submits the job (jar/executable etc.) and configuration to the JobTracker which then assumes the responsibility of distributing the software / configuration to the slaves, scheduling tasks and monitoring them, providing status and diagnostic information to the job-client.
Ans: The MapReduce framework operates exclusively on pairs, that is, the framework views the input to the job as a set of pairs and produces a set of pairs as the output of the job, conceivably of different types.
See the flow mentioned below
(input) -> map -> -> combine/sorting -> -> reduce -> (output)
Ans: the key and value classes have to be serialized by the framework. To make them serializable Hadoop provides a Writable interface. As you know from the java itself that the key of the Map should be comparable, hence the key has to implement one more interface Writable Comparable.
Ans: We will count the words in all the input file flow as below
input Assume there are two files each having a sentence Hello World Hello World (In file 1) Hello World Hello World (In file 2)
Mapper : There would be each mapper for the a file For the given sample input the first map output:
< Hello, 1>
< World, 1>
< Hello, 1>
< World, 1>
The second map output:
< Hello, 1>
< World, 1>
< Hello, 1>
< World, 1>
Combiner/Sorting (This is done for each individual map) So output looks like this The output of the first map:
< Hello, 2>
< World, 2>
The output of the second map:
< Hello, 2>
< World, 2>
Reducer : It sums up the above output and generates the output as below
< Hello, 4>
< World, 4>
Final output would look like
Hello 4 times
World 4 times
Ans: Maps are the individual tasks that transform input records into intermediate records. The transformed intermediate records do not need to be of the same type as the input records. A given input pair may map to zero or many output pairs.
Ans: An InputSplit is a logical representation of a unit (A chunk) of input work for a map task; e.g., a file name and a byte range within that file to process or a row set in a text file.
Ans: The InputFormat is responsible for enumerate (itemise) the InputSplits, and producing a RecordReader which will turn those logical work units into actual physical input records.
Ans: Generally mapper implementation is specified in the Job itself.
Ans: The Mapper itself is instantiated in the running job, and will be passed a MapContext object which it can use to configure itself.
Ans: The Mapper contains the run() method, which call its own setup() method only once, it also call a map() method for each input and finally calls it cleanup() method. All above methods you can override in your code.
Ans: If you do not override any methods (leaving even map as-is), it will act as the identity function, emitting each input record as a separate output.
Ans: The Context object allows the mapper to interact with the rest of the Hadoop system. It Includes configuration data for the job, as well as interfaces which allow it to emit output.
Ans: You can set arbitrary (key, value) pairs of configuration data in your Job, e.g. with
Job.getConfiguration().set("myKey", "myVal"), and then retrieve this data in your mapper with
Context.getConfiguration().get("myKey"). This kind of functionality is typically done in the Mapper's setup() method
Ans: The Mapper.run() method then calls map(KeyInType, ValInType, Context) for each key/value pair in the InputSplit for that task
Ans: The output of the Mapper are sorted and Partitions will be created for the output. Number of partition depends on the number of reducer.
Ans: Users can control which keys (and hence records) go to which Reducer by implementing a custom Partitioned.
Ans: It is an optional component or class, and can be specify via Job.setCombinerClass(ClassName), to perform local aggregation of the intermediate outputs, which helps to cut down the amount of data transferred from the Mapper to the Reducer.
Ans: The number of maps is usually driven by the total size of the inputs, that is, the total number of blocks of the input files.
Generally it is around 10-100 maps per-node. Task setup takes awhile, so it is best if the maps take at least a minute to execute.
Suppose, if you expect 10TB of input data and have a block size of 128MB, you'll end up with 82,000 maps, to control the number of block you can use the mapreduce.job.maps parameter (which only provides a hint to the framework). Ultimately, the number of tasks is controlled by the number of splits returned by the InputFormat.getSplits() method (which you can override).
Ans: Reducer reduces a set of intermediate values which share a key to a (usually smaller) set of values.
The number of reduces for the job is set by the user via Job.setNumReduceTasks(int).
Ans: The API of Reducer is very similar to that of Mapper, there's a run() method that receives a Context containing the job's configuration as well as interfacing methods that return data from the reducer itself back to the framework. The run() method calls setup() once, reduce() once for each key associated with the reduce task, and cleanup() once at the end. Each of these methods can access the job's configuration data by using Context.getConfiguration().
As in Mapper, any or all of these methods can be overridden with custom implementations. If none of these methods are overridden, the default reducer operation is the identity function; values are passed through without further processing.
The heart of Reducer is its reduce() method. This is called once per key; the second argument is an Iterable which returns all the values associated with that key.
Ans: Shuffle, Sort and Reduce.
Ans: Input to the Reducer is the sorted output of the mappers. In this phase the framework fetches the relevant partition of the output of all the mappers, via HTTP.
Ans: The framework groups Reducer inputs by keys (since different mappers may have output the same key) in this stage. The shuffle and sort phases occur simultaneously; while map-outputs are being fetched they are merged (It is similar to merge-sort).
Ans: In this phase the reduce(MapOutKeyType, Iterable, Context) method is called for each pair in the grouped inputs. The output of the reduce task is typically written to the FileSystem via Context.write (ReduceOutKeyType, ReduceOutValType). Applications can use the Context to report progress, set application-level status messages and update Counters, or just indicate that they are alive. The output of the Reducer is not sorted.
Ans: The right number of reduces seems to be 0.95 or 1.75 multiplied by
( * mapreduce.tasktracker.reduce.tasks.maximum).
With 0.95 all of the reduces can launch immediately and start transfering map outputs as the maps finish. With 1.75 the faster nodes will finish their first round of reduces and launch a second wave of reduces doing a much better job of load balancing. Increasing the number of reduces increases the framework overhead, but increases load balancing and lowers the cost of failures.
I hope this blog may help a few of you to learn and crack any big data Hadoop interview questions asked by top companies. As per the Gartner report, almost 62% of the big companies use big data Hadoop software to control an enormous amount of data. So you can expect the highest number of Hadoop job openings with huge salary offers. To learn more about big data Hadoop, please visit our HKR website.