Homo-Adminus Blog

Yet another small note about Cloudera Hadoop Distribution 4.3.

This time I needed to deploy some custom JAR files to our Hive cluster so that we wouldn’t need to do “ADD JAR” commands in every Hive job (especially useful when using HiveServer API).

Here is the process of adding a custom SerDE or a UDF jar to your Cloudera Hadoop cluster:

• First, we have built our JSON SerDe and got a json-serde-1.1.6.jar file.
• To make this file available to Hive CLI tools, we need to copy it to /usr/lib/hive/lib on every server in the cluster (I have prepared an rpm package to do just that).
• To make sure Hive map-reduce jobs would be able to read/write JSON tables, we needed to copy our JAR file to /usr/lib/hadoop/lib directory on all task tracker servers in the cluster (the same rpm does that).
• And last, really important step: To make sure your TaskTracker servers know about the new jar, you need to restart your tasktracker services (we use Cloudera Manager, so that was just a few mouse clicks ;-))

And this is it for today.

Just a short note to myself and others who need to add LZO support for CDH 4.3.

First of all, you need to build hadoop-lzo. Since CDH 4.3 uses hadoop 2.0, most of the forks of hadoop-lzo project fail to compile against new libraries. After some digging I’ve found the original twitter hadoop-lzo branch to be the most maintained and it works perfectly with hadoop 2.0. So, download it, install pre-requisites, build it.

I have built it for us as an RPM, you can check out the spec file here (it depends on some other packages from that repo, but you should get the idea and should be able to modify the script to build on vanilla Redhat linux w/o additional packages). Another option would be to take a look at Cloudera’s GPL Extras repository and their lzo packages and documentation.

After you have built and installed your LZO libraries, you should be able to use them with HBase without any additional configuration. To test HBase support for LZO compression you could use the following command:

You’re looking for that last line to say SUCCESS. If it fails, it means you did something wrong and it will tell you what that is.

Now, if you want to use LZO for map-reduce jobs, you need to make a few changes in your /etc/hadoop/conf/core-site.xml config file. If you manage your configuration yourself, just add the following to your configuration file:

If you’re managing your configuration with Cloudera Manager, you need to do the following:

1. Go to your map-reduce service
2. Click “Configuration” and select “View and Edit“
3. In the list on the left select “Gateway (Default)” and “Compression“
4. Add two items to the list of compression codecs: com.hadoop.compression.lzo.LzoCodec and com.hadoop.compression.lzo.LzoCodec
5. Open “Service Wide” => “Advanced” in the list on the left
6. Add the following configuration to your “MapReduce Service Configuration Safety Valve for mapred-site.xml” section:

   1 2 3 4

   <property>   <name>io.compression.codec.lzo.class</name>   <value>com.hadoop.compression.lzo.LzoCodec</value> </property>

7. Click “Save Changes“
8. Restart your map-reduce cluster with updated configuration

Now you should be able to use LZO in your map-reduce, hive and pig jobs.