Hadoop installation directory

Question

Which directory is Hadoop installed in Cloudera distribution? Is it in /usr/bin/hadoop?

[cloudera@quickstart opt]$ which hadoop
/usr/bin/hadoop

I know the software packages are to be installed inside /opt/ directory.

What does Apache recommend?

Answer 1

Actually if you use parcels for Cloudera CDH (which is recommended way to install it) it goes under /opt/cloudera/parcels/CDH which is in turn symlink to actual CDH parcel. Under this directory you will find structure very similar to what open source Apache Hadoop normally hase under / .

More, if you browse under / you will find Hadoop normal binaries being symlinks which point to CDH active parcel through alternatives. For example check /usr/bin/hadoop .

This installation strategy brings Cloudera with ability to activate new CDH in a second and it looks like normal Apache Hadoop structure. Under CDH parcel directory you will find 2-layer binaries structure. For example in contrast to normal Apache Hadoop with entry point /usr/bin/hadoop under CDH you will have /opt/cloudera/parcels/CDH/bin/hadoop entry point and actual script /opt/cloudera/parcels/CDH/lib/hadoop/bin/hadoop .

Libraries are placed in more proprietary manner. All JARS are under /opt/cloudera/parcels/CDH/jars and appropriate JARS in common library directory or product specific ones are just symlinks. For example check hbase specific libs: /opt/cloudera/parcels/CDH/lib/hbase/lib/ .