Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Hadoop Operations
Search
Marc Cluet
June 09, 2013
Technology
130
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Hadoop Operations
Marc Cluet
June 09, 2013
More Decks by Marc Cluet
See All by Marc Cluet
FOSDEM'14 - Autoscaling Best Practices
lynxman
1
130
A metadata ocean in Chef and Puppet
lynxman
0
67
Rackspace Hack Night - Vagrant & Packer
lynxman
0
160
Innovation in the Cloud - Rackspace Zurich Event
lynxman
0
110
Introduction to DevOps - Rackspace Tech Night
lynxman
1
88
Introduction To Hadoop
lynxman
1
120
SSH That Wonderful Thing
lynxman
1
100
Networking & DNS 101
lynxman
0
110
Juju and Puppet - Rapid Harmonious Deployment
lynxman
0
110
Other Decks in Technology
See All in Technology
AWSとGitHub Actionsの責任境界と 組織で安全に使用する取り組み
nealle
1
220
When Token Pruning is Worse than Random: Understanding Visual Token Information in VLLMs
sansantech
PRO
0
160
AI時代の「OAuth認証」にどう物申すか?(OAuth/OIDC Numa (Immersion) Workshop 2026)
oidfj
PRO
0
350
dbt と Snowflake と tag
kevinrobot34
2
300
みてねにおけるAI-DLC導入活動とAIドリブン開発の現在地/JAWS-UG AI-DLC #2
isaoshimizu
2
290
エージェント化するAI:現在地とその先に起きる変化 / AI as Agents: The Current State and the Changes Ahead
ks91
PRO
0
160
8bit CPU 2026
koba789
7
3k
500名弱規模の組織のPythonプロジェクト(dbt) をどう管理するか?
hiracky16
0
370
AI駆動開発をチームに根付かせる - 「1行も書かない」チームがHarnessを育てた1年 -
kenichirokimura
6
3.3k
Introduction to Sansan, inc / Sansan Global Development Center, Inc.
sansan33
PRO
0
3.2k
サーバー常駐型の 簡易障害調査AI エージェントを作ってみた話
masayoshi
1
380
DMMブックスのNext.js化を加速させるAI活用 / Migrating DMM Books to Next.js with AI
kentarom
0
200
Featured
See All Featured
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
47
8.3k
Tips & Tricks on How to Get Your First Job In Tech
honzajavorek
1
710
Marketing to machines
jonoalderson
1
5.7k
jQuery: Nuts, Bolts and Bling
dougneiner
66
8.6k
Abbi's Birthday
coloredviolet
3
9.6k
Odyssey Design
rkendrick25
PRO
2
780
16th Malabo Montpellier Forum Presentation
akademiya2063
PRO
0
360
Optimizing for Happiness
mojombo
378
71k
Color Theory Basics | Prateek | Gurzu
gurzu
0
440
Navigating the Design Leadership Dip - Product Design Week Design Leaders+ Conference 2024
apolaine
2
410
How People are Using Generative and Agentic AI to Supercharge Their Products, Projects, Services and Value Streams Today
helenjbeal
1
290
Learning to Love Humans: Emotional Interface Design
aarron
275
41k
Transcript
Marc Cluet – Lynx Consultants How Hadoop Works
What we’ll cover? ¡ Understand Hadoop in detail ¡
See how Hadoop works operationally ¡ Be able to start asking the right questions from your data Lynx Consultants © 2013
Hadoop Distributions ¡ Cloudera CDH ¡ Hortonworks ¡
MapR Lynx Consultants © 2013
Hadoop Components ¡ HDFS ¡ Hbase ¡ MapRed
¡ YARN Lynx Consultants © 2013
Hadoop Components ¡ HDFS § Hadoop Distributed File System
§ Everything sits on top of it § Has 3 copies by default of every block ¡ Hbase ¡ MapRed ¡ YARN Lynx Consultants © 2013
Hadoop Components ¡ HDFS ¡ Hbase § Hadoop
Schemaless Database § Key value Store § Sits on top of HDFS ¡ MapRed ¡ YARN Lynx Consultants © 2013
Hadoop Components ¡ HDFS ¡ Hbase ¡ MapRed
§ Hadoop Map/Reduce § Non-‐pluggable, archaic § Requires HDFS for temp storage ¡ YARN Lynx Consultants © 2013
Hadoop Components ¡ HDFS ¡ Hbase ¡ MapRed
¡ YARN § Hadoop Map/Reduce version 2.0 § Pluggable, you can add your own § Fast and not so much memory hungry Lynx Consultants © 2013
Hadoop Component Breakdown ¡ All these components divide themselves in
§ client/server § master/slave scenarios ¡ We will now check each individual component breakdown Lynx Consultants © 2013
Hadoop Components Breakdown ¡ HDFS § Master Namenode
▪ Keeps track of all file allocation on Datanodes ▪ Rebalances data if one of the namenodes goes down ▪ Is Rack aware § Secondary Namenode ▪ Does cleanup services for the namenode ▪ Not necessarily two different servers § Datanode ▪ Stores the data ▪ Good to have not RAID disks for extra I/O speed Lynx Consultants © 2013
Hadoop Components Breakdown ¡ HDFS § How to access
▪ Client can connect with hadoop client to hdfs://namenode:8020 ▪ Supports all basic Unix commands § Configuration files ▪ /etc/hadoop/conf/core-‐site.xml ▪ Defines major configuration as hdfs namenode and default parameters ▪ /etc/hadoop/conf/hdfs-‐site.xml ▪ Defines configuration specific to namenode or datanode on file locations ▪ /etc/hadoop/conf/slaves ▪ Defines the list of servers that are available in this cluster Lynx Consultants © 2013
Hadoop Components Breakdown ¡ Hbase § Master ▪
Controls the Hbase cluster, knows where the data is allocated and provides a client listening socket using Thrift and/or a RESTful API § Regionserver ▪ Hbase node, stores some of the information in one of the regions, it’d be equivalent to sharding § Thrift / REST ▪ Interface to connect to HBase Lynx Consultants © 2013
Hadoop Components Breakdown ¡ Hbase § How to access
▪ Through the Hbase client (using Thrift) ▪ Through the RESTful API § Configuration files ▪ /etc/hbase/conf/hbase-‐site.xml ▪ Defines all the basic configuration for accessing hbase ▪ /etc/hbase/conf/hbase-‐policy.xml ▪ Defines all the security (ACL) and all the hbase memory tweaks ▪ /etc/hbase/conf/regionservers ▪ List all the regionservers available to this cluster Lynx Consultants © 2013
Hadoop Components Breakdown ¡ MapRed § JobTracker ▪
Creates the Map/Reduce jobs ▪ Stores all the intermediate data ▪ Keeps track of all the previous results through the HistoryServer § TaskTracker ▪ Executed Tasks related to the Map/Reduce job ▪ Very CPU and memory intensive ▪ Stores intermediate results which then are pushed to JobTracker Lynx Consultants © 2013
Hadoop Components Breakdown ¡ MapRed § How to access
▪ Through the Hadoop Client ▪ Through any MapRed client like Pig or Hive ▪ Own Java code § Configuration files ▪ /etc/hadoop/conf/mapred-‐site.xml ▪ Defines how to contact this MapRed Cluster ▪ /etc/hadoop/conf/mapred-‐queue-‐acls.xml ▪ Defines ACL structure for accessing MapRed, normally not necessary ▪ /etc/hadoop/conf/slaves ▪ Defines the list of TaskTrackers in this cluster Lynx Consultants © 2013
Hadoop Components Breakdown ¡ YARN § Same structure as
MapRed (lives on top of it) § Configuration files ▪ /etc/hadoop/conf/yarn-‐site.xml ▪ All required configuration for YARN Lynx Consultants © 2013
Hadoop Cluster Breakdown ¡ Namenode Server § HDFS Namenode
§ Hbase Master ¡ Secondary Namenode Server § HDFS Secondary Namenode ¡ JobTracker Server § MapRed JobTracker § MapRed History Server Lynx Consultants © 2013
Hadoop Cluster Breakdown ¡ Datanode Server § HDFS Datanode
§ Hbase RegionServer § MapRed TaskTracker Lynx Consultants © 2013
Hadoop Hardware Requirements ¡ Namenode Server § Redundant power
supplies § RAID1 Drives § Enough memory (16Gb) ¡ Secondary Namenode Server § Almost none Lynx Consultants © 2013
Hadoop Hardware Requirements ¡ Jobtracker Server § Redundant power
supplies § RAID1 Drives § Enough memory (16Gb) ¡ Datanode Server § Lots of cheap disk (no RAID) § Lots of memory (32Gb) § Lots of CPU Lynx Consultants © 2013
Hadoop Default Ports ¡ HDFS § 8020: HDFS Namenode
§ 50010: HDFS Datanode FS transfer ¡ MapRed § No defaults ¡ Hbase § 60010: Master § 60020: Regionserver Lynx Consultants © 2013
Hadoop HDFS Workflow Lynx Consultants © 2013
Hadoop MapRed Workflow Lynx Consultants © 2013
Hadoop MapRed Workflow Lynx Consultants © 2013
Flume ¡ Transports streams of data from point A to
point B ¡ Source § Where the data is read from ¡ Channel § How the data is buffered ¡ Sink § Where the data is written Lynx Consultants © 2013
Flume ¡ Flume is fault tolerant ¡ Sources are
pointer kept § With some exceptions, but most sources are in a known state ¡ Channels can be fault tolerant § Channel written to disk can recover from where it left ¡ Sinks can be redundant § More than one sink for the same data § Data is serialised and deduplicated using AVRO Lynx Consultants © 2013
Flume Lynx Consultants © 2013
Flume ¡ Configuration files § /etc/flume-‐ng/conf/flume.conf ▪ Defines
the agent configuration with source, channel, sink Lynx Consultants © 2013
Flume Lynx Consultants © 2013
Hadoop Recommended Reads Lynx Consultants © 2013
Hadoop References ¡ Hadoop § http://hadoop.apache.org/docs/stable/cluster_setup.html § http://rc.cloudera.com/cdh/4/hadoop/hadoop-‐yarn/hadoop-‐yarn-‐site/
ClusterSetup.html § http://pig.apache.org/docs/r0.7.0/setup.html § http://wiki.apache.org/hadoop/NameNodeFailover ¡ Hbase § http://hbase.apache.org/book/book.html ¡ Flume § http://archive.cloudera.com/cdh4/cdh/4/flume-‐ng/ FlumeUserGuide.html Lynx Consultants © 2013
Questions? Lynx Consultants © 2013