| 1 | = 2011-09-12 = |
| 2 | |
| 3 | == Hadoop == |
| 4 | |
| 5 | * [http://www.hortonworks.com/set-up-apache-hadoop-in-minutes-with-rpms/ Set Up Apache Hadoop in Minutes with RPMs] |
| 6 | * 今天看到這篇文章,回想起幾天前的 [http://ftp.twaren.net/Unix/Web/apache//hadoop/common/hadoop-0.20.204.0/ Hadoop 0.20.204.0 Release],回過頭去查目錄內容,才赫然發現裡面包含了 RPM 跟 DEB 套件檔,看樣子未來套件安裝又多了一種選擇。 |
| 7 | * [http://www.hortonworks.com/an-introduction-to-hdfs-federation/ An Introduction to HDFS Federation | Hortonworks] |
| 8 | * [Image(http://www.hortonworks.com/wp-content/uploads/2011/08/HDFS_Federation_PNG2.png)]] |
| 9 | * 看起來是多個 !NameNode 形成多個 Storage Pool !NameSpace 的作法,比較特別的是 !DataNode 是共用的!與過去比較常看到的多組 !NameNode 各自配對應的 !DataNode 作法不同。 |
| 10 | * [http://www.hortonworks.com/nextgen-mapreduce-hits-apache-hadoop-mainline/ Preparing for Next Release of Apache Hadoop] |
| 11 | * 先前也看過 [http://developer.yahoo.com/blogs/hadoop/posts/2011/02/mapreduce-nextgen/ Yahoo Developer Network 的文章在談次世代 MapReduce],本來沒有很懂差異點。今天再看了一下[http://www.slideshare.net/hortonworks/nextgen-apache-hadoop-mapreduce 投影片],才明白是要把原本 !JobTracker 做的事情分成多個角色來處理,包含:!ResourceManager、!NodeManager與!ApplicationMaster。 |
| 12 | * [[Image(http://ydn.zenfs.com/blogs/22/MapReduce_NextGen.jpg)]] |