Hortonworks HDPCD Dumps

The mapper output (intermediate data) is stored on the Local file system (NOT HDFS) of each individual mapper nodes. This is typically a temporary directory location which can be setup in config by the hadoop administrator. The intermediate data is cleaned up after the Hadoop Job completes.

Note:

* Join Algorithms in MapReduce

A) Reduce-side join

B) Map-side join

C) In-memory join

/ Striped Striped variant variant

/ Memcached variant

* Which join to use?

/ In-memory join > map-side join > reduce-side join

/ Limitations of each?

In-memory join: memory

Map-side join: sort order and partitioning

Reduce-side join: general purpose

The fundamental idea of MRv2 is to split up the two major functionalities of the JobTracker, resource management and job scheduling/monitoring, into separate daemons. The idea is to have a global ResourceManager (RM) and per-application ApplicationMaster (AM). An application is either a single job in the classical sense of Map-Reduce jobs or a DAG of jobs.

Note:

The central goal of YARN is to clearly separate two things that are unfortunately smushed together in current Hadoop, specifically in (mainly) JobTracker:

/ Monitoring the status of the cluster with respect to which nodes have which resources available. Under YARN, this will be global.

/ Managing the parallelization execution of any specific job. Under YARN, this will be done separately for each job.

Multiple reducers and total ordering

If your sort job runs with multiple reducers (either because mapreduce.job.reduces in mapred-site.xml has been set to a number larger than 1, or because you’ve used the -r option to specify the number of reducers on the command-line), then by default Hadoop will use the HashPartitioner to distribute records across the reducers. Use of the HashPartitioner means that you can’t concatenate your output files to create a single sorted output file. To do this you’ll need total ordering,

Oozie workflow is a collection of actions (i.e. Hadoop Map/Reduce jobs, Pig jobs) arranged in a control dependency DAG (Direct Acyclic Graph), specifying a sequence of actions execution. This graph is specified in hPDL (a XML Process Definition Language).

hPDL is a fairly compact language, using a limited amount of flow control and action nodes. Control nodes define the flow of execution and include beginning and end of a workflow (start, end and fail nodes) and mechanisms to control the workflow execution path ( decision, fork and join nodes).

Workflow definitions

Currently running workflow instances, including instance states and variables

See 3) below.

Limitations of Mapreduce – where not to use Mapreduce

While very powerful and applicable to a wide variety of problems, MapReduce is not the answer to every problem. Here are some problems I found where MapReudce is not suited and some papers that address the limitations of MapReuce.

1. Computation depends on previously computed values

If the computation of a value depends on previously computed values, then MapReduce cannot be used. One good example is the Fibonacci series where each value is summation of the previous two values. i.e., f(k+2) = f(k+1) + f(k). Also, if the data set is small enough to be computed on a single machine, then it is better to do it as a single reduce(map(data)) operation rather than going through the entire map reduce process.

2. Full-text indexing or ad hoc searching

The index generated in the Map step is one dimensional, and the Reduce step must not generate a large amount of data or there will be a serious performance degradation. For example, CouchDB’s MapReduce may not be a good fit for full-text indexing or ad hoc searching. This is a problem better suited for a tool such as Lucene.

3. Algorithms depend on shared global state

Solutions to many interesting problems in text processing do not require global synchronization. As a result, they can be expressed naturally in MapReduce, since map and reduce tasks run independently and in isolation. However, there are many examples of algorithms that depend crucially on the existence of shared global state during processing, making them difficult to implement in MapReduce (since the single opportunity for global synchronization in MapReduce is the barrier between the map and reduce phases of processing)

You can use your reducer code as a combiner if the operation performed is commutative and associative.

Use Apache HBase when you need random, realtime read/write access to your Big Data.

Note: This project's goal is the hosting of very large tables -- billions of rows X millions of columns -- atop clusters of commodity hardware. Apache HBase is an open-source, distributed, versioned, column-oriented store modeled after Google's Bigtable: A Distributed Storage System for Structured Data by Chang et al. Just as Bigtable leverages the distributed data storage provided by the Google File System, Apache HBase provides Bigtable-like capabilities on top of Hadoop and HDFS.

Features

Linear and modular scalability.

Strictly consistent reads and writes.

Automatic and configurable sharding of tables

Automatic failover support between RegionServers.

Convenient base classes for backing Hadoop MapReduce jobs with Apache HBase tables.

Easy to use Java API for client access.

Block cache and Bloom Filters for real-time queries.

Query predicate push down via server side Filters

Thrift gateway and a REST-ful Web service that supports XML, Protobuf, and binary data encoding options

Extensible jruby-based (JIRB) shell

Support for exporting metrics via the Hadoop metrics subsystem to files or Ganglia; or via JMX