Dynamic error capture with Clojure

I’ve recently been trying to improve parts of uSwitch.com’s Clojure pricing code for the last few days. Mostly this has been to add error handling that provides a form of graceful degradation.

The majority of the application deals with pricing the various consumer Energy plans that people can pick from. Although we try and make sure everything validates before it hits the site we’ve found a few weird problems that weren’t anticipated. This can be bad when errors prevent the other results from being returned successfully.

Ideally, we’d like to be able to capture both the errors and the plans we couldn’t price; clients can use the data we have and know what’s missing.

I’m sure I had read an interview in which Rich Hickey said how Clojure’s dynamism helped deal with error handling in a clean and safe way. I can’t find a reference (so it’s possible I made that up), but, I then was looking through The Joy of Clojure and found this:

“there are two directions for handling errors. The first … refers to the passive handling of exceptions bubbling outward from inner functions. But built on Clojure’s dynamic Var binding is a more active mode of error handling, where handlers are pushed into inner functions.”

The code below shows a trivialised version of the problem- a calculation function that might raise an error part-way through the collection.

Instead, we can use binding and an error handling function to dynamically handle the problem and push this down the stack.

Note that we’re using declare to define our error-handler Var, and, that we have to mark it as :dynamic as we’re going to be dynamically binding it.

The next stage is to progress from just returning nil and to capture the errors whilst the records are being processed. We can use Atoms to hold the state and then append more information as we flow through.

The above example introduces an errors atom that will capture the responses as we map across the sequence. But it definitely now feels a little icky.

  1. We use do to both add the error to our list of errors, and return nil.
  2. Because we’ve now introduced side-effects we must use doall to realise the sequence whilst the handle-error function is rebound.

I have to say, I’m not too sure whether I prefer this more dynamic approach or a more traditional try...catch form.

The dynamic behaviour is definitely very cool, and the authors of The Joy of Clojure say that “using dynamic scope via binding is the preferred way to handle recoverable errors in a context-sensitive manner.” so I’m very inclined to stick with it. However, I think I’ve managed to make it less nice: what originally looked like a neat use of closure to capture errors in a safe way now seems less good? I’d love to hear what more experienced Clojurists make of it.

Other references:

  1. http://kotka.de/blog/2010/05/Didyouknow_IV.html
  2. http://kotka.de/blog/2009/11/TamingtheBound_Seq.html
  3. http://cemerick.com/2009/11/03/be-mindful-of-clojures-binding/

Using partial to tidy Clojure tests

I was reading through getwoven’s infer code on GitHub and noticed a nice use of partial to tidy tests.

Frequently there’ll be a function under test that will be provided multiple inputs with perhaps only a single parameter value changing per example. Using partial you can build a function that closes over it’s variable arguments letting you tidy up the call to the single (varying) parameter.

The example I found was in infer’s sparse_classification_test.clj but I’ve written a quick gist to demonstrate the point:

Recognising SSL Requests in Rack with Custom Headers

For a while now I’ve been working on replacing bits of uSwitch.com’s .NET heritage with some new Clojure and Ruby blood. A problem with one of the new applications came up when we needed it to recognise (and behave differently) when an SSL secured request was made because of non-standard headers (although I’m yet to successfully Google the RFC where HTTP_X_FORWARDED_PROTO is defined :).

Polyglot Deployment

New applications are deployed on different machines but proxied by the old application servers- letting us mount new applications inside the current website. Although the servers run IIS and ASP.NET they’re configured much like nginx’s proxy_pass module (for example). You set the matching URL and the URL you’d like to proxy to. uSwitch.com also use hardware load balancers in front of the application and web servers.

Rack Scheme Detection

Recent releases of Rack include a ssl? method on the Request that return true if the request is secure. This, internally, calls scheme to check whether any of the HTTP headers that specify an HTTPS scheme are set (the snippet from Rack’s current code is below).

Unfortunately, this didn’t recognise our requests as being secure. The hardware load-balancer was setting a custom header that wasn’t anything Rack would expect. The quickest (and perhaps dirty?) solution- a gem, rack-scheme-detect, that allows additional mappings to be provided- the rest of Rack will continue to work (and all of the other middleware and apps above it).

The following snippet shows how it can be used inside a classic Sinatra application to configure the additional mapping.

Of course we’ll be making sure we also set one of the standard headers. But, for anyone who needs a quick solution when that’s not immediately possible, hopefully rack-scheme-detect might save some time.

Multiple connections with Hive

We’ve been using Hadoop and Hive at Forward for just over a year now. We have lots of batch jobs written in Ruby (using our Mandy library) and a couple in Clojure (using my fork of cascading-clojure). This post covers a little of how things hang together following a discussion on the Hive mailing list around opening multiple connections to the Hive service.

Hive has some internal thread safety problems (see HIVE-80) that meant we had to do something a little boutique with HAProxy given our usage. This article covers a little about how Hive fits into some of our batch and ad-hoc processing and how we put it under HAProxy.

Background

Hive is used for both programmatic jobs and interactively by analysts. Interactive queries come via a web-based product, called CardWall, that was built for our search agency. In both cases, behind the scenes, we have another application that handles scheduling jobs, retrying when things fail (and alerting if it’s unlikely to work thereafter).

Batch Scheduling

Internally, this is how things (roughly) look:

Batch processing

The Scheduler is responsible for knowing which jobs need to run and queuing work items. Within the batch processing application an Agent will pick up an item and check whether it’s dependencies are available (we use this to integrate with all kinds of other reporting systems, FTP sites etc.). If things aren’t ready it will place the item back on the queue with a delay (we use Beanstalkd which makes this all very easy).

Enter Hive

You’ll notice from the first screenshot we have some Hive jobs that are queued. When we first started using Hive we started using a single instance of the HiveServer service started as follows: $HIVE_HOME/bin/hive --service hiveserver. Mostly this was using a Ruby Gem called RBHive.

This worked fine when we were using it with a single connection, but, we ran into trouble when writing apps that would connect to Hive as part of our batch processing system.

Originally there was no job affinity, workers would pick up any piece of work and attempt it, and with nearly 20 workers this could heavily load the Hive server. Over time we introduced different queues to handle different jobs, allowing us to allocate a chunk of agents to sensitive resources (like Hive connections).

Multiple Hive Servers

We have a couple of machines that run multiple HiveServer instances. We use Upstart to manage these processes (so have hive1.conf, hive2.conf… in our /etc/init directory). Each config file specifies a port to open the service on (10001 in the example below).

These are then load-balanced with HAProxy, the config file we use for that is below

Using Clojure Protocols for Java Interop

Clojure's protocols were a feature I'd not used much of until I started working on a little pet project: clj-hector (a library for accessing the Cassandra datastore).

Protocols provide a means for attaching type-based polymorphic functions to types. In that regard they're somewhat similar to Java's interfaces (indeed Java can interoperate with Clojure through the protocol-generated types although it's not something I've tried). Here's the example of their use from the Clojure website:

Protocols go deeper though: they can be used with already defined types (think attaching behaviour through mixing modules into objects with Ruby's include). It's for this reason that they've been a particularly nice abstraction for building adapters upon: glue that translates Java objects to Clojure structures when doing interop.

Here's a snippet from a bit of the code that converts Hector query results into maps:

QueryResultImpl contains results for queries. In one instance, this is RowsImpl which contains Row results (which in turn are converted by other parts of the extend-protocol statement in the real code). The really cool part of this is that dispatch happens polymorphically: there's no need to write code to check types, dispatch (and call recursively down the tree). Instead, you map declaratively by type and build a set of statements about how to convert. Clojure does the rest.

Pretty neat.

Processing null character terminated text files in Hadoop with a NullByteTextInputFormat

It was Forward's first hackday last Friday. Tom Hall and I worked on some collaborative filtering using Mahout and Hadoop to process some data collected from Twitter.

I'd been collecting statuses through their streaming API (using some code from a project called Twidoop). After wrestling with formats for a little while we realised that each JSON record was separated by a null character (all bits zeroed). There didn't seem to be an easy way to change the terminating character for the TextInputFormat that we'd normally use so we wrote our own InputFormat and I'm posting it here for future reference (and anyone else looking for a Hadoop InputFormat to process similar files).

We added it to my fork of cascading-clojure but it will work in isolation.

Managing Deployment SSH Keys

At Forward the number of virtual machines we're deploying to is increasing steadily; on EC2 alone we have over 30 at the moment. Managing authentication to those servers was becoming more time consuming.

Previously we'd used a specific user with a password that was shared between those that needed to access the machines and keeping that up to date was often unreliable; machines would not be updated with the new password, and everybody would have to be told of the new password.

We wanted something better so we're now using a git repository to sync public keys.

It's easy to manage, easy to add new keys, and easy to track changes. We can manage permissions to the repository, remove keys when necessary, and it's very easy to make sure all machines are constantly up to date.

To do this we have a repository that contains a set of user.pub public key files, copied directly from the user's ~/.ssh/id_dsa.pub file (for example).

Machine images have the GitHub signature already accepted and a clone of the repository. A simple Bash script then executes regularly via cron, pulling any changes and updating the authorized_keys file.

It's been working pretty well so far, much easier than before!

Virtualisation, Levels of Abstraction and Thinking Infrastructure

It’s 2010 and although Hoverboards are a little closer I still travel mostly by bus. In contrast, the way we’re using computing at Forward definitely seems to be mirroring a wider progression towards utility computing. What’s interesting is how this is actually achieved in two subtly different ways.

Firstly, let’s take the classic example: services deployed onto a distributed and external virtual machine environment managed via an API. We use this for some of our most important systems (managing millions and millions of transactions a day) and for a couple of great reasons: Amazon’s EC2 service offers us a level of distribution that would be very expensive (and time consuming) to build ourselves. As George mentioned in his post about our continual deployment process we make use of 4 geographically distributed compute zones.

This is virtualisation, but, at a pretty low-level. I could fire up a couple of EC2 nodes and do what I liked with them: deploy a bunch of Sinatra apps on Passenger, spin-up a temporary Hadoop cluster to do some heavy lifting, or perhaps do some video encoding.

Using the classic EC2 model I (as a consumer of the service) need to understand how to make use of the services that I then deploy. Of course, Amazon make it a little easier with pre-bundled AMI’s that contain pre-installed packages, but, this still needs me to be aware: pick the right CPU architecture, find the AMI with the version of RabbitMQ I’m after etc.

A lot of other ‘cloud’ providers (think Joyent Accelerators, Rackspace’s CloudServers SliceHost etc.) are very similar. Although you can programmatically control instances, only pay for the time you use them for etc., you’re still thinking at a relatively low systems level.

Amazon’s Elastic MapReduce Service is an example of a higher-level virtualised abstraction: I don’t care what’s going on underneath (although I have to pay depending on the capacity I want to give it). I submit my job and wait for the reply.

Heroku is another great example of this kind of higher-level service: deploy your code straight from the command-line, dynamically allocate resources etc. I don’t have to worry about a caching infrastructure- it’s built in. My application just needs to be a good HTTP citizen and things just work. Bliss.

Recently we made an investment in some dedicated hardware to replace the existing virtualised infrastructure that ran our Hadoop cluster. As alluded to in the original MapReduce paper: both the implementation and development model encourage a general model for large-scale data processing. Squint at your problem for long enough and it’ll probably fit into the MapReduce model. It’s not always that pleasant (or productive) to do that so there are a number of higher-level abstraction atop the map/reduce data flows to choose from: Cascading, Pig, and Hive are some good examples for Hadoop, Google also have their Sawzall paper.

Underneath all of that, however, is still a general platform for distributed computation: each layer builds on the previous.

MapReduce (and distributed storage), therefore, provide a kind of virtualisation albeit at a higher-level of abstraction to your average virtual machine. We’re consolidating workloads onto the same infrastructure.

We’re slowly moving more and more of our batch processing onto this infrastructure and (consequently) simplifying the way we deal with substantial growth. Batch processing large data is becoming part of our core infrastructure and, most importantly, is then able to be re-used by other parts of the business.

It feels like there are two different kinds of virtualisation at play here: Amazon EC2 (which offers raw compute power) and platforms like Hadoop which can provide a higher-level utility to a number of consumers. Naturally the former often provides the infrastructure to provide the latter (Elastic MapReduce being a good example).

Perhaps more significant is the progression towards even higher-levels of abstraction.

Google’s Jeff Dean gave a talk late last year (sorry, can’t find the link) about the next generation infrastructure that Google was building: the infrastructure was becoming intelligent.

Rather than worrying about how to deploy an application to get the best from it, and by building it upon some core higher-level services, the system could adapt to meet constraints. Need requests to the EU to be served within 1ms? The system could ensure data is replicated to a rack in a specific region. Need a batch to be finished by 9:00am? The system could ensure enough compute resources are allocated.

Amazon’s Elastic Load Balancing service includes an Auto Scale feature: set conditions that describe when instances should be added or removed and it will automatically respond. That’s great, but, I’d rather think in terms of application requirements. It’s a subtle shift in emphasis, much like the move from an imperative to declarative style.

I have no doubt that virtualisation has been profoundly significant. But, what really excites me is the move towards higher-level services that let me deploy into a set of infrastructure that can adapt to meet my requirements. It sounds as crazy as Hoverboards, but, it doesn’t feel that distant a reality.

Processing XML in Hadoop

This is something that I’ve been battling with for a while and tonight, whilst not able to sleep, I finally found a way to crack it! The answer: use an XmlInputFormat from a related project. Now for the story.

We process lots of XML documents every day and some of them are pretty large: over 8 gigabytes uncompressed. Each. Yikes!

We’d made significant performance gains by switching the old REXML SAX parser to libxml. But we’ve suffered from random segfaults on our production servers (seemingly caused by garbage collecting bad objects). Besides, this still was running at nearly an hour for the largest reports.

Hadoop did seem to offer XML processing: the general advice was to use Hadoops’s StreamXmlRecordReader which can be accessed through using the StreamInputFormat.

This seemed to have weird behaviour with our reports (which often don’t have line-endings): lines would be duplicated and processing would jump to 100%+ complete. All a little fishy.

Hadoop Input Formats and Record Readers

The InputFormat in Hadoop does a couple of things. Most significantly, it provides the Splits that form the chunks that are sent to discrete Mappers.

Splits form the rough boundary of the data to be processed by an individual Mapper. The FileInputFormat (and it’s subclasses) generate splits on overall file size. Of course, it’s unlikely that all individual Records (a Key and Value that are passed to each Map invocation) lie neatly within these splits: records will often cross splits. The RecordReader, therefore, must handle this

… on whom lies the responsibilty to respect record-boundaries and present a record-oriented view of the logical InputSplit to the individual task.

In short, a RecordReader could scan from the start of a split for the start of it’s record. It may then continue to read past the end of it’s split to find the end. The InputSplit only contains details about the offsets from the underlying file: data is still accessed through the streams.

It seemed like the StreamXmlRecordReader was skipping around the underlying InputStream too much; reading records it wasn’t entitled to read. I tried my best at trying to understand the code but it was written a long while ago and is pretty cryptic to my limited brain.

I started trying to rewrite the code from scratch but it became pretty hairy very quickly. Take a look at the implementation of next() in LineRecordReader to see what I mean.

Mahout to the Rescue

After a little searching around on GitHub I found another XmlInputFormat courtesy of the Lucene sub-project: Mahout.

I’m happy to say it appears to work. I’ve just run a quick test on our 30 node cluster (via my VPN) and it processed the 8 gig file in about 10 minutes. Not bad.

For anyone trying to process XML with Hadoop: try Mahout’s XmlInputFormat.

MapReduce with Hadoop and Ruby

The quantity of data we analyse at Forward every day has grown nearly ten-fold in just over 18 months. Today we run almost all of our automated daily processing on Hadoop. The vast majority of that analysis is automated using Ruby and our open-source gem: Mandy.

In this post I’d like to cover a little background to MapReduce and Hadoop and a light introduction to writing a word count in Ruby with Mandy that can run on Hadoop.

Mandy’s aim is to make MapReduce with Hadoop as easy as possible: providing a simple structure for writing code, and commands that make it easier to run Hadoop jobs.

Since I left ThoughtWorks and joined Forward in 2008 I’ve spent most of my time working on the internal systems we use to work with our clients, track our ads, and analyse data. Data has become truly fundamental to our business.

I think it’s worth putting the volume of our processing in numbers: we track around 9 million clicks a day across our campaigns- averaged out across a day that’s over 100 every second (it actually peaks up to around 1400 a second).

 

We automate the analysis, storage, and processing of around 80GB (and growing) of data every day. In addition, further ad-hoc analysis is performed through the use of a related project: Hive (more on this in a later blog post).

We’d be hard pressed to work with this volume of data in a reasonable fashion, or be able to change what analysis we run, so quickly if it wasn’t for Hadoop and it’s associated projects.

Hadoop is an open-source Apache project that provides “open-source software for reliable, scalable, distributed computing”. It’s written in Java and, although relatively easy to get into, still falls foul of Java development cycle problems: they’re just too long. My colleague (Andy Kent) started writing something we could prototype our jobs in Ruby quickly, and then re-implement in Java once we understood things better. However, this quickly turned into our platform of choice.

To give a taster of where we’re going, here’s the code needed to run a distributed word count:

And how to run it?

mandy-hadoop wordcount.rb hdfs-dir/war-and-peace.txt hdfs-dir/wordcount-output

The above example comes from a lab Andy and I ran a few months ago at our office. All code (and some readmes) are available on GitHub.

What is Hadoop?

The Apache Hadoop project actually contains a number of sub-projects. MapReduce probably represents the core- a framework that provides a simple distributed processing model, which when combined with HDFS provides a great way to distribute work across a cluster of hardware. Of course, it goes without saying that it is very closely based upon the Google paper.

Parallelising Problems

Google’s MapReduce paper explains how, by adopting a functional style, programs can be automatically parallelised across a large cluster of machines. The programming model is simple to grasp and, although it takes a little getting used to, can be used to express problems that are seemingly difficult to parallelise.

Significantly, working with larger data sets (both storage and processing) can be solved by growing the cluster’s capacity. In the last 6 months we’ve grown our cluster from 3 ex-development machines to 25 nodes and taken our HDFS storage from just over 1TB to nearly 40TB.

An Example

Consider an example: perform a subtraction across two sets of items. That is

{1,2,3} - {1,2} = {3}

. A naive implementation might compare every item for the first set against those in the second set.

numbers = [1,2,3]
[1,2].each { |n| numbers.delete(n) }

This is likely to be problematic when we get to large sets as the time taken will grow very fast (m*n). Given an initial set of 10,000,000 items, finding the distinct items from a second site of even 1,000 items would result in 10 billion operations.

However, because of the way the data flows through a MapReduce process, the data can be partitioned across a number of machines such that each machine performs a smaller subset of operations with the result then combined to produce a final output.

It’s this flow that makes it possible to ‘re-phrase’ our solution to the problem above and have it operate across a distributed cluster of machines.

Overall Data Flow

The Google paper and a Wikipedia article provide more detail on how things fit together, but here’s the rough flow.

Map and reduce functions are combined into one or more ‘Jobs’ (almost all of our analysis is performed by jobs that pass their output to further jobs). The general flow can be seen to be made of the following steps:

  1. Input data read
  2. Data is split
  3. Map function called for each record
  4. Shuffle/sort
  5. Sorted data is merged
  6. Reduce function called
  7. Output (for each reducer) written to a split
  8. Your application works on a join of the splits

First, the input data is read and partitioned into splits that are processed by the Map function. Each map function then receives a whole ‘record’. For text files, these records will be a whole line (marked by a carriage return/line feed at the end).

Output from the Map function is written behind-the-scenes and a shuffle/sort performed. This prepares the data for the reduce phase (which needs all values for the same key to be processed together). Once the data has been sorted and distributed to all the reducers it is then merged to produce the final input to the reduce.

Map

The first function that must be implemented is the map function. This takes a series of key/value pairs, and can emit zero or more pairs of key/value outputs. Ruby already provides a method that works similarly: Enumerable#map.

For example:

names = %w(Paul George Andy Mike)
names.map {|name| {name.size => name}} # => [{4=>"Paul"}, {6=>"George"}, {4=>"Andy"}, {4=>"Mike"}]

The map above calculates the length of each of the names, and emits a list of key/value pairs (the length and names). C# has a similar method in ConvertAll that I mentioned in a previous blog post.

Reduce

The reduce function is called once for each unique key across the output from the map phase. It also produces zero or more key value pairs.

For example, let’s say we wanted to find the frequency of all word lengths, our reduce function would look a little like this:

def reduce(key, values)
  {key => values.size}
end

Behind the scenes, the output from each invocation would be folded together to produce a final output.

To explore more about how the whole process combines to produce an output let’s now turn to writing some code.

Code: Word Count in Mandy

As I’ve mentioned previously we use Ruby and Mandy to run most of our automated analysis. This is an open-source Gem we’ve published which wraps some of the common Hadoop command-line tools, and provides some infrastructure to make it easier to write MapReduce jobs.

Installing Mandy

The code is hosted on GitHub, but you can just as easily install it from Gem Cutter using the following:

$ sudo gem install gemcutter
$ gem tumble
$ sudo gem install mandy

… or

$ sudo gem install mandy --source http://gemcutter.org

If you run the mandy command you may now see:

You need to set the HADOOP_HOME environment variable to point to your hadoop install    :(
Try setting 'export HADOOP_HOME=/my/hadoop/path' in your ~/.profile maybe?

You’ll need to make sure you have a 0.20.x install of Hadoop, with HADOOP_HOME set and $HADOOP_HOME/bin in your path. Once that’s done, run the command again and you should see:

You are running Mandy!
========================

Using Hadoop 0.20.1 located at /Users/pingles/Tools/hadoop-0.20.1

Available Mandy Commands
------------------------
mandy-map       Run a map task reading on STDIN and writing to STDOUT
mandy-local     Run a Map/Reduce task locally without requiring hadoop
mandy-rm        remove a file or directory from HDFS
mandy-hadoop    Run a Map/Reduce task on hadoop using the provided cluster config
mandy-install   Installs the Mandy Rubygem on several hosts via ssh.
mandy-reduce    Run a reduce task reading on STDIN and writing to STDOUT
mandy-put       upload a file into HDFS

If you do, you’re all set.

Writing a MapReduce Job

As you may have seen earlier to write a Mandy job you can start off with as little as:

Of course, you may be interested in writing some tests for your functions too. Well, you can do that quite easily with the Mandy::TestRunner. Our first test for the Word Count mapper might be as follows:

Our implementation for this might look as follows:

Of course, our real mapper needs to do more than just tokenise the string. It also needs to make sure we downcase and remove any other characters that we want to ignore (numbers, grammatical marks etc.). The extra specs for these are in the mandy-lab GitHub repository but are relatively straightforward to understand.

Remember that each map works on a single line of text. We emit a value of 1 for each word we find which will then be shuffled/sorted and partitioned by the Hadoop machinery so that a subset of data is sent to each reducer. Each reducer will, however, receive all the values for that key. All we need to do in our reducer is sum all the values (all the 1’s) and we’ll have a count. First, our spec:

And our implementation:

In fact, Mandy has a number of built-in reducers that can be used (so you don’t need to re-implement this). To re-use the Sum-ming reducer replace the reduce statement above to reduce(Mandy::Reducers::SumReducer).

Our mandy-lab repository includes some sample text files you can run the examples on. Assuming you’re in the root of the repo you can then run the following to perform the wordcount across Alice in Wonderland locally:

$ mandy-local word_count.rb input/alice.txt output
Running Word Count...
/Users/pingles/.../output/1-word-count

If you open the /Users/pingles/.../output/1-word-count file you should see a list of words and their frequency across the document set.

Running the Job on a Cluster

The mandy-local command just uses some shell commands to imitate the MapReduce process. If you try running the same command over a larger dataset you’ll come unstuck. So let’s see how we run the same example on a real Hadoop cluster. Note, that for this to work you’ll need to either have a Hadoop cluster or Hadoop running locally in pseudo-distributed mode.

If you have a cluster up and running, you’ll need to write a small XML configuration file to provide the HDFS and JobTracker connection info. By default the mandy-hadoop command will look for a file called cluster.xml in your working directory. It should look a little like this:

<?xml version="1.0" encoding="UTF-8"?>
<configuration>
  <property>
    <name>fs.default.name</name>
    <value>hdfs://datanode:9000/</value>
  </property>
  <property>
    <name>mapred.job.tracker</name>
    <value>jobtracker:9001</value>
  </property>
  <property>
    <name>hadoop.job.ugi</name>
    <value>user,supergroup</value>
  </property>
</configuration>

Once that’s saved, you need to copy your text file to HDFS so that it can be distributed across the cluster ready for processing. As an aside: HDFS shards data in 64MB blocks across the nodes. This provides redundancy to help mitigate in the case of machine failure. It also means the job tracker (the node which orchestrates the processing) can move the computation close to the data and avoid copying large amounts of data unnecessarily.

To copy a file up to HDFS, you can run the following command

$ mandy-put my-local-file.txt word-count-example/remote-file.txt

You’re now ready to run your processing on the cluster. To re-run the same word count code on a real Hadoop cluster, you change the mandy-local command to mandy-hadoop as follows:

$ mandy-hadoop word_count.rb word-count-example/remote-file.txt word-count-example/output

Once that’s run you should see some output from Hadoop telling you the job has started, and where you can track it’s progress (via the HTTP web console):

Packaging code for distribution...
Loading Mandy scripts...

Submitting Job: [1] Word Count...
Job ID: job_200912291552_2336
Kill Command: mandy-kill job_200912291552_2336 -c /Users/pingles/.../cluster.xml
Tracking URL: http://jobtracker:50030/jobdetails.jsp?jobid=job_200912291552_2336

word-count-example/output/1-word-count

Cleaning up...
Completed Successfully!

If you now take a look inside ./word-count-example/output/1-word-count you should see a few part-xxx files. These contain the output from each reducer. Here’s the output from a reducer that ran during my job (running across War and Peace):

abandoned   54
abandoning  26
abandonment 14
abandons    1
abate   2
abbes   1
abdomens    2
abduction   3
able    107

Wrap-up

That’s about as much as I can cover in an introductory post for Mandy and Hadoop. I’ll try and follow this up with a few more to show how you can pass parameters from the command-line to Mandy jobs, how to serialise data between jobs, and how to tackle jobs in a map and reduce stylee.

As always, please feel free to email me or comment if you have any questions, or (even better) anything you’d like me to cover in a future post.