Achieving Multiple Active Namenodes in Hadoop HDFS

Acheving Multiple Active Namenodes in Hadoop HDFS

Step 1. Download the Apache Hadoop Source Code

git clone https://github.com/apache/hadoop

git clone https://github.com/apache/hadoop
cd hadoop
git checkout rel/release-3.2.1

Need to check if head hash as prefix is as follows

b3cbbb4

Step 2. Apply the Patch and Re-build it

*Number of files to be modified: 30, Number of lines of Java code to be added: 7,000

Step 3. Prepare the Node (minimal option)

2 Active Namenodes
1 Datanode

*The followings are not needed more: QJM, Journal Nodes (2k+1), Zookeeper, Standby Namenode, Observer Namenode, etc.

Step 4. Deploy the Hadoop (one-touch installation without human-intervention)

Step 5. Launch the Hadoop (one-touch launching using only-one script without complex scripts)

Step 6. Upload the files (via HDFS API like Hive/Presto does)

Using HDFS API is recommended. WebGUI or CLI can be also used.

Step 7. Check the files replicated over all Namenodes

Publications

2021 IEEE International IOT, Electronics and Mechatronics Conference

https://ieeexplore.ieee.org/Xplore/home.jsp