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.

Today, just like many times before, I needed to configure a monitoring server for MySQL using Cacti and awesome Percona Monitoring Templates. The only difference was that this time I wanted to get it to run with 1 min resolution (using ganglia and graphite, both with 10 sec resolution, for all the rest of our monitoring in Swiftype really spoiled me!). And that’s where the usual pain in the ass Cacti configuration gets really amplified by the million things you need to change to make it work. So, this is a short checklist post for those who need to configure a Cacti server with 1 minute resolution and setup Percona Monitoring Plugins on it.

Read the rest of this entry →

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.

So, after a few weeks of looking for a new job I’m really excited to start my journey in a young, but very ambitious startup called Swiftype which is focused on developing a technology for private site search, that could be used on everything from small blogs to large product sites. The company is growing really fast and I’m going to lead all the work on infrastructure, build the ops team and hope to get a chance to do some coding along the way.

Stay tuned – I really hope to finally get a chance to do more blogging this year. 🙂

As of today I’m no longer working for LivingSocial and I’m looking for the next thing to work on. Since my family is in Toronto and I have an apartment (mortgage) here, I’m not looking to relocate and currently looking for something remote (I have 7+ years of remote work experience) or something local in Toronto.

For more information on my background, please check my Github profile, my linkedin profile or the resume section on this blog. If you need to contact me, feel free to use any channels listed on the contacts page.

Update: After a few initial interviews I’d like to update this post with a bit more details on what I’m looking for in the new position.

Though even more I’m interested in being able to make strategic technical decisions for an operations team and apply my experience and knowledge for solving challenging tasks with a dedicated team of ops engineers. This could be anything from a tech ops team lead role (in a medium/large companies) to a director of technical operations (in a small-to-medium sized startups).

Update: Ok, I’ve found a new job – I work for Swiftype now!