Archive

Author Archive

Apache Solr real-time live index updates at scale with Apache Hadoop

Episode # 22 of the podcast was a talk with Patrick Hunt

We talked about the new work that has gone into Apache Solr (upstream) that allows it to work on Apache Hadoop. Solr has support for writing and reading its index and transaction log files to the HDFS distributed filesystem. This does not use Hadoop Map-Reduce to process Solr data, rather it only uses the HDFS filesystem for index and transaction log file storage. https://cwiki.apache.org/confluence/display/solr/Running+Solr+on+HDFS

We also talked about Solr Cloud and how the sharding features allow Solr to scale with a Hadoop cluster https://cwiki.apache.org/confluence/display/solr/SolrCloud.

Apache Solr includes the ability to set up a cluster of Solr servers that combines fault tolerance and high availability. Called SolrCloud, these capabilities provide distributed indexing and search capabilities, supporting the following features:

  • Central configuration for the entire cluster
  • Automatic load balancing and fail-over for queries
  • ZooKeeper integration for cluster coordination and configuration.

SolrCloud is flexible distributed search and indexing, without a master node to allocate nodes, shards and replicas. Instead, Solr uses ZooKeeper to manage these locations, depending on configuration files and schemas. Documents can be sent to any server and ZooKeeper will figure it out.

Patrick introduced me to Morphlines (part of the Cloudera Development Kit for Hadoop) http://cloudera.github.io/cdk/docs/current/cdk-morphlines/index.html

Cloudera Morphlines is an open source framework that reduces the time and skills necessary to build and change Hadoop ETL stream processing applications that extract, transform and load data into Apache Solr, HBase, HDFS, Enterprise Data Warehouses, or Analytic Online Dashboards. Want to build or facilitate ETL jobs without programming and without substantial MapReduce skills? Get the job done with a minimum amount of fuss and support costs? Here is how to get started.

A morphline is a rich configuration file that makes it easy to define a transformation chain that consumes any kind of data from any kind of data source, processes the data and loads the results into a Hadoop component. It replaces Java programming with simple configuration steps, and correspondingly reduces the cost and integration effort associated with developing and maintaining custom ETL projects.

Morphlines is a library, embeddable in any Java codebase. A morphline is an in-memory container of transformation commands. Commands are plugins to a morphline that perform tasks such as loading, parsing, transforming, or otherwise processing a single record. A record is an in-memory data structure of name-value pairs with optional blob attachments or POJO attachments. The framework is extensible and integrates existing functionality and third party systems in a straightforward manner.

The morphline commands were developed as part of Cloudera Search. Morphlines power ETL data flows from Flume and MapReduce and HBase into Apache Solr. Flume covers the real time case, whereas MapReduce covers the batch processing case. Since the launch of Cloudera Search morphline development graduated into the Cloudera Development Kit(CDK) in order to make the technology accessible to a wider range of users and products, beyond Search. The CDK is a set of libraries, tools, examples, and documentation focused on making it easier to build systems on top of the Hadoop ecosystem. The CDK is hosted on GitHub and encourages involvement by the community. For example, morphlines could be embedded into Crunch, HBase, Impala, Pig, Hive, or Sqoop. Let us know where you want to take it!

Morphlines can be seen as an evolution of Unix pipelines where the data model is generalized to work with streams of generic records, including arbitrary binary payloads. A morphline is an efficient way to consume records (e.g. Flume events, HDFS files, RDBMS tables or Avro objects), turn them into a stream of records, and pipe the stream of records through a set of easily configurable transformations on the way to a target application such as Solr, for example as outlined in the following figure:

In this figure, a Flume Source receives syslog events and sends them to a Flume Morphline Sink, which converts each Flume event to a record and pipes it into a readLine command. The readLine command extracts the log line and pipes it into a grok command. The grok command uses regular expression pattern matching to extract some substrings of the line. It pipes the resulting structured record into the loadSolr command. Finally, the loadSolr command loads the record into Solr, typically a SolrCloud. In the process, raw data or semi-structured data is transformed into structured data according to application modelling requirements.

The Morphline framework ships with a set of frequently used high level transformation and I/O commands that can be combined in application specific ways. The plugin system allows the adding of new transformations and I/O commands and integrates existing functionality and third party systems in a straightforward manner.

This integration enables rapid Hadoop ETL application prototyping, complex stream and event processing in real time, flexible log file analysis, integration of multiple heterogeneous input schemas and file formats, as well as reuse of ETL logic building blocks across Hadoop ETL applications.

The CDK ships an efficient runtime that compiles a morphline on the fly. The runtime executes all commands of a given morphline in the same thread. Piping a record from one command to another implies just a cheap Java method call. In particular, there are no queues, no handoffs among threads, no context switches and no serialization between commands, which minimizes performance overheads.

Morphlines manipulate continuous or arbitrarily large streams of records. A command transforms a record into zero or more records. The data model can be described as follows: A record is a set of named fields where each field has an ordered list of one or more values. A value can be any Java Object. That is, a record is essentially a hash table where each hash table entry contains a String key and a list of Java Objects as values. Note that a field can have multiple values and any two records need not use common field names. This flexible data model corresponds exactly to the characteristics of the Solr/Lucene data model.

Not only structured data, but also binary data can be passed into and processed by a morphline. By convention, a record can contain an optional field named _attachment_body, which can be a Java java.io.InputStream or Java byte[]. Optionally, such binary input data can be characterized in more detail by setting the fields named _attachment_mimetype (such as “application/pdf”) and _attachment_charset (such as “UTF-8″) and _attachment_name (such as “cars.pdf”), which assists in detecting and parsing the data type. This is similar to the way email works.

This generic data model is useful to support a wide range of applications. For example, the Apache Flume Morphline Solr Sink embeds the morphline library and executes a morphline to convert flume events into morphline records and load them into Solr. This sink fills the body of the Flume event into the _attachment_body field of the morphline record, as well as copies the headers of the Flume event into record fields of the same name. As another example, the Mappers of the MapReduceIndexerTool fill the Java java.io.InputStream referring to the currently processed HDFS file into the _attachment_body field of the morphline record. The Mappers of the MapReduceIndexerTool also fill metadata about the HDFS file into record fields, such as the file’s name, path, size, last modified time, etc. This way a morphline can act on all data received from Flume and HDFS. As yet another example, the Morphline Lily HBase Indexer fills a HBase Result Java POJO into the _attachment_body field of the morphline record. This way morphline commands such as extractHBaseCells can extract data from HBase updates and correspondingly update a Solr index.

We also talked a good deal about Apache Zookeeper and some of the history back from when Zookeeper was originally at Yahoo! and Patrick’s experience since then. To hear everything that Patrick had to say please subscribe to the podcast.

 

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

 

Reporting Metrics to Apache Kafka and Monitoring with Consumers

April 18, 2014 2 comments

Apache Kafka has been used for some time now by organizations to consume not only all of the data within its infrastructure from an application perspective but also the server statistics of the running applications and infrastructure.  Apache Kafka is great for this.

Coda Hale’s metrics’s has become a leading way to instrument your JVM applications capturing what the application is doing and reporting it to different servers like Graphite, Ganglia, Riemann and other systems.  The main problem with this is the tight coupling between your application (and infrastructure) metrics with how you are charting, trending and alerting on them.  Now lets insert Apache Kafka to-do the decoupling which it does best.

kafka1

The systems sending data and the systems reading the data become decoupled through the Kafka brokers.

producer_consumer

 

Now, once this decoupling happens it allows us to plug in new systems (producers) to send metrics and then have multiple different systems consuming them.  This means that not only can you monitor and alert on the metrics but also (and at the same time) do real time analysis or analysis over the larger dataset consumed off somewhere else.

So, how does this all work? Well, we created a Metrics reporter https://github.com/stealthly/metrics-kafka that you can use within your own applications and any application supporting Coda Hale’s metrics but now reporting to a Kafka topic.  We will be creating more producers (such as scrapping cpu, mem, disk, etc) and adding them to this project.  To use this setup a Metrics reporter like you would normally but do so like this:

import ly.stealth.kafka.metrics.KafkaReporter
val producer = KafkaReporter.builder(registry, kafkaConnection, topic).build()
producer.report()
It was developed in Java so you don’t have to include the Scala dependencies if you are not using Scala. The source code for the reporter is available here: https://github.com/stealthly/metrics-kafka/tree/master/metrics/src/main/java/ly/stealth/kafka/metrics.
We also setup an example of how to read the produced metrics with Riemann. Riemann monitors distributed systems and provides low-latency, transient shared state for systems with many moving parts.  There is a KafkaRiemman consumer that reads from Kafka and posts to Riemann https://github.com/stealthly/metrics-kafka/blob/master/riemann/src/main/scala/ly/stealth/kafka/riemann/RiemannMetricsConsumer.scala.
1) Install Vagrant http://www.vagrantup.com/
2) Install Virtual Box https://www.virtualbox.org/
git clone https://github.com/stealthly/metrics-kafka.git
cd metrics-kafka
vagrant up
Once the virtual machines are launched go to http://192.168.86.55:4567/.  We setup a sample dashboard. Fom your command prompt run the test cases from the command prompt (after vms have launched).
./gradlew test
And the dashboard will populate automatically.
Screen Shot 2014-04-18 at 12.04.13 PM
You can modify what is showing up in Riemann and learn how to use this from the test cases https://github.com/stealthly/metrics-kafka/blob/master/riemann/src/test/scala/ly/stealth/kafka/metrics/RiemannMetricsConsumerSpec.scala.
 
/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

 

Categories: Open Source Projects

Higher Availability, Increased Scale and Enhanced Security on Apache HBase

April 3, 2014 Leave a comment

Episode #21 of the Podcast was a talk with Michael Stack, Lars Hofhansl and Andrew Purtell.

Having these guests from the Apache HBase PMC allowed us to talk about HBase 0.96, 0.98, some use cases, HBaseCon and 1.0.

The highlights from 0.96 where around stability and longer term scale (moving all internal data exchange and persistence to protobufs).

0.98 introduced some exciting new security features and a new HFile format with both encryption at rest and cell level security labels.

HBaseCon has all new speakers and new use cases with new and familiar faces listening onward. A must attend if you can make it.

1.0 is focusing on SLA and more inmemorry database features and general cleanup.

Listen into the podcast and all of what they talked about together.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

 

Categories: HBase, Podcast

Beyond MapReduce and Apache Hadoop 2.X with Bikas Saha and Arun Murthy

April 2, 2014 Leave a comment

Episode 20 of the podcast was with Bikas Saha and Arun Murthy.

When I spoke with Arun a year or so a go YARN was NextGen Hadoop and there have been a lot of updates, work done and production experience since.

Besides Yahoo! other multi thousand node clusters have been and are running in production with YARN. These clusters have shown 2x capacity throughput which resulted in reduced cost for hardware (and in some cases being able to shut down co-los) while still gaining performance improvements overall to previous clusters of Hadoop 1.X.

I got to hear about some of what is in 2.4 and coming in 2.5 of Hadoop:

  • Application timeline server repository and api for application specific metrics (Tez, Spark, Whatever).
    • web service API to put and get with some aggregation.
    • plugable nosql store (hbase, accumulo) to scale it.
  • Preemption capacity scheduler.
  • Multiple resource support (CPU, RAM and Disk).
  • Labels tag nodes with labels can be labeled however so some windows and some linux and ask for resources with only those labels with ACLS.
  • Hypervisor support as a key part of the topology.
  • Hoya generalize for YARN (game changer) and now proposed as Slider to the Apache incubator.

We talked about Tez which provides complex DAGs of queries to translate what you want to-do on Hadoop without the work arounds for making it have to run in MapReduce.  MapReduce was not designed to be re-workable out side of the parts of the Job it gave you for Map, Split, Shuffle, Combine, Reduce, Etc and Tez is more expressible exposing a DAG API.

PigHiveQueryOnMR

Now becomes with Tez:

PigHiveQueryOnTez

 

There were also some updates on Hive v13 coming out with sub queries, low latency queries (through Tez), high precision decimal points and more!

Subscribe to the podcast and listen to all of what Bikas and Arun had to say.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

 

 

Big Data with Apache Accumulo Preserving Security with Open Source

March 13, 2014 Leave a comment

Episode 19 of the podcast was a talk with Adam Fuchs.

Adam talked about Apache Accumulo which is a system built for doing random i/o with peta bytes of data.

Distributing the computation to the data with cell level security is where Accumulo really shines.

Accumulo provides a richer data model than simple key-value stores, but is not a fully relational database. Data is represented as key-value pairs, where the key and value are comprised of the following elements:

Key Value
Row ID Column Timestamp
Family Qualifier Visibility

All elements of the Key and the Value are represented as byte arrays except for Timestamp, which is a Long. Accumulo sorts keys by element and lexicographically in ascending order. Timestamps are sorted in descending order so that later versions of the same Key appear first in a sequential scan. Tables consist of a set of sorted key-value pairs.

Accumulo stores data in tables, which are partitioned into tablets. Tablets are partitioned on row boundaries so that all of the columns and values for a particular row are found together within the same tablet. The Master assigns Tablets to one TabletServer at a time. This enables row-level transactions to take place without using distributed locking or some other complicated synchronization mechanism. As clients insert and query data, and as machines are added and removed from the cluster, the Master migrates tablets to ensure they remain available and that the ingest and query load is balanced across the cluster.

images/data_distribution.png
images/failure_handling.png
Subscribe to the Podcast and here all of what Adam had to say.
You can get started using Apache Accumulo with our development environment https://github.com/stealthly/hdp-accumulo
/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

XML to Avro Conversion

February 26, 2014 1 comment

We all know what XML is right?  Just in case not, no problem here is what it is all about.  


<root>
<node>5</node>
</root>

Now, what the computer really needs is the number five and some context around it. In XML you (human and computer) can see how it represents context to five. Now lets say instead you have a business XML document like FPML


<FpML xmlns="http://www.fpml.org/2007/FpML-4-4" xmlns:fpml="http://www.fpml.org/2007/FpML-4-4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="4-4" xsi:schemaLocation="http://www.fpml.org/2007/FpML-4-4 ../fpml-main-4-4.xsd http://www.w3.org/2000/09/xmldsig# ../xmldsig-core-schema.xsd" xsi:type="RequestTradeConfirmation">
<!--  start of distinct  -->
<strike>
<strikePrice>32.00</strikePrice>
</strike>
<numberOfOptions>150000</numberOfOptions>
<optionEntitlement>1.00</optionEntitlement>
<equityPremium>
<payerPartyReference href="party2"/>
<receiverPartyReference href="party1"/>
<paymentAmount>
<currency>EUR</currency>
<amount>405000</amount>
</paymentAmount>
<paymentDate>
<unadjustedDate>2001-07-17Z</unadjustedDate>
<dateAdjustments>
<businessDayConvention>NONE</businessDayConvention>
</dateAdjustments>
</paymentDate>
<pricePerOption>
<currency>EUR</currency>
<amount>2.70</amount>
</pricePerOption>
</equityPremium>
</equityOption>
<calculationAgent>
<calculationAgentPartyReference href="party1"/>
</calculationAgent>
<documentation>
<masterAgreement>
<masterAgreementType>ISDA2002</masterAgreementType>
</masterAgreement>
<contractualDefinitions>ISDA2002Equity</contractualDefinitions>
<!--
 populate credit support document with correct value 
-->
<creditSupportDocument>TODO</creditSupportDocument>
</documentation>
<governingLaw>GBEN</governingLaw>
</trade>
<party id="party1">
<partyId>Party A</partyId>
</party>
<party id="party2">
<partyId>Party B</partyId>
</party>
</FpML>

That is a lot of extra unnecessary data points. Now lets look at this using Apache Avro.

With Avro, the context and the values are separated. This means the schema/structure of what the information is does not get stored or streamed over and over and over and over (and over) again.

The Avro schema is hashed. So the data structure only holds the value and the computer understands the fingerprint (the hash) of the schema and can retrieve the schema using the fingerprint.

0x d7a8fbb307d7809469ca9abcb0082e4f8d5651e46d3cdb762d02d0bf37c9e592

This type of implementation is pretty typical in the data space.

When you do this you can reduce your data between 20%-80%. When I tell folks this they immediately ask, “why such a large gap of unknowns”. The answer is because not every XML is created the same. But that is the problem because you are duplicating the information the computer needs to understand the data. XML is nice for humans to read, sure … but that is not optimized for the computer.

Here is a converter we are working on https://github.com/stealthly/xml-avro to help get folks off of XML and onto lower cost, open source systems. This allows you to keep parts of your systems (specifically the domain business code) using the XML and not having to be changed (risk mitigation) but store and stream the data with less overhead (optimize budget).

/*******************************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC

http://www.stealth.ly

Twitter: @allthingshadoop
********************************************/

Impala and SQL on Hadoop

February 22, 2014 Leave a comment

The origins of Impala can be found in F1 – The Fault-Tolerant Distributed RDBMS Supporting Google’s Ad Business.

One of many differences between MapReduce and Impala is in Impala the intermediate data moves from process to process directly instead of storing it on HDFS for processes to get at the data needed for processing.  This provides a HUGE performance advantage and doing so while consuming few cluster resources.   Less hardware to-do more!

impala

There are many advantages to this approach over alternative approaches for querying Hadoop data, including::

  • Thanks to local processing on data nodes, network bottlenecks are avoided.
  • A single, open, and unified metadata store can be utilized.
  • Costly data format conversion is unnecessary and thus no overhead is incurred.
  • All data is immediately query-able, with no delays for ETL.
  • All hardware is utilized for Impala queries as well as for MapReduce.
  • Only a single machine pool is needed to scale.

We encourage you to read the documentation for further exploration!

There are still transformation steps required to optimize the queries but Impala can help to-do this for you with Parquet file format.  Better compression and optimized runtime performance is realized using the ParquetFormat though many other file types are supported.

This and a whole lot more was discussed with Marcel Kornacker the Cloudera Architect behind Impala on Episode 18 of the All Things Hadoop Podcast.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

Apache Kafka HTTP Producer and Consumer

January 16, 2014 Leave a comment

Big Data Open Source Security LLC has released an Apache Kafka HTTP endpoint for producing and consuming data with Apache Kafka brokers https://github.com/stealthly/dropwizard-kafka-http.

This project creates an Apache Kafka HTTP endpoint for producing and consuming messages.

Running Project

1) Install Vagrant (http://www.vagrantup.com/)
2) Install Virtual Box (https://www.virtualbox.org/)
3) git clone https://github.com/stealthly/dropwizard-kafka-http
4) cd dropwizard-kafka-http
5) vagrant up

Then produce and consume messages over HTTP e.g.

curl -d "topic=http&message=hello&key=0" "http://192.168.22.10:8080/message"

and

curl "http://192.168.22.10:8080/message?topic=http"

Your original message produced from the consumer:

[{"topic":"http","key":"0","message":"hello","partition":0,"offset":0}]

* Zookeeper will be running on 192.168.22.5
* dropwizard-kafka-http is built cleanly before Zookeeper installs in `vagrant/zk.sh`
* Broker One is running on 192.168.22.10
* dropwizard-kafka-http launches on 192.168.22.10 after the Kafka Broker starts in `vagrant/broker.sh`

If you want you can login to the machines using `vagrant ssh zookeeper` and `vagrant ssh brokerOne`.

API Methods

Produce messages

POST /message?topic=$topic&async={true|false}&message=$message0&key=$key0,
message=$message1,key=$key1, …

Produces messages on topic.

Parameters:
topic – required topic name
async – optional true|false value indicating should the producer be async
key – required key text. Multiple values are possible
message – required message text. Multiple values are possible

Note: passed keys count should match messages count.

Errors:
400 – wrong parameters passed

Consume messages

GET /message?topic=$topic&timeout=$timeout
Returns consumed messages available on topic.

Parameters:
topic – required topic name
timeout – optional timeout in ms

Errors:
400 – wrong parameters passed

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/
Categories: Open Source Projects

Developing Apache Kafka Producers and Consumers

December 21, 2013 Leave a comment

I gave a presentation recently on Real-time streaming and data pipelines with Apache Kafka.

A correction in the talk (~ 22 minutes in) : I said that you have to have all your topic data fit on one server.  That is not true, you can’t span logs so you have to have all of your data for a partition fit on one server.  Kafka will spread your partitions around for you within topics.

For that presentation I put together sample code for producing and consuming with an Apache Kafka broker using Scala.

To get up and running, use vagrant.

1) Install Vagrant http://www.vagrantup.com/
2) Install Virtual Box https://www.virtualbox.org/

git clone https://github.com/stealthly/scala-kafka
cd scala-kafka
vagrant up
./sbt test

Your entry point is the test file

/**
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *    http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package ly.stealth.testing

import org.specs2.mutable._
import java.util.UUID
import kafka.consumer._
import kafka.producer._
import kafka.utils._
import kafka.akka._
import akka.actor.{Actor, Props, ActorSystem}
import akka.routing.RoundRobinRouter

class KafkaSpec extends Specification with Logging {

  "Simple Producer and Consumer" should {
    "send string to broker and consume that string back" in {
      val testMessage = UUID.randomUUID().toString
      val testTopic = UUID.randomUUID().toString
      val groupId_1 = UUID.randomUUID().toString

      var testStatus = false

      info("starting sample broker testing")
      val producer = new KafkaProducer(testTopic,"192.168.86.10:9092")
      producer.sendString(testMessage)

      val consumer = new KafkaConsumer(testTopic,groupId_1,"192.168.86.5:2181")

      def exec(binaryObject: Array[Byte]) = {
        val message = new String(binaryObject)
        info("testMessage = " + testMessage + " and consumed message = " + message)
        testMessage must_== message
        consumer.close()
        testStatus = true
      }

      info("KafkaSpec is waiting some seconds")
      consumer.read(exec)
      info("KafkaSpec consumed")

      testStatus must beTrue // we need to get to this point but a failure in exec will fail the test
    }

    "send string to broker and consume that string back in different consumer groups" in {
      val testMessage = UUID.randomUUID().toString
      val testTopic = UUID.randomUUID().toString
      val groupId_1 = UUID.randomUUID().toString
      val groupId_2 = UUID.randomUUID().toString

      var testStatus1 = false
      var testStatus2 = false

      info("starting sample broker testing")
      val producer = new KafkaProducer(testTopic,"192.168.86.10:9092")
      producer.sendString(testMessage)

      val consumer1 = new KafkaConsumer(testTopic,groupId_1,"192.168.86.5:2181")

      def exec1(binaryObject: Array[Byte]) = {
        val message1 = new String(binaryObject)
        info("testMessage 1 = " + testMessage + " and consumed message 1 = " + message1)
        testMessage must_== message1
        consumer1.close()
        testStatus1 = true
      }

      info("KafkaSpec : consumer 1 - is waiting some seconds")
      consumer1.read(exec1)
      info("KafkaSpec : consumer 1 - consumed")

      val consumer2 = new KafkaConsumer(testTopic,groupId_2,"192.168.86.5:2181")

      def exec2(binaryObject: Array[Byte]) = {
        val message2 = new String(binaryObject)
        info("testMessage 2 = " + testMessage + " and consumed message 2 = " + message2)
        testMessage must_== message2
        consumer2.close()
        testStatus2 = true
      }

      info("KafkaSpec : consumer 2 - is waiting some seconds")
      consumer2.read(exec2)
      info("KafkaSpec : consumer 2 - consumed")

      testStatus2 must beTrue // we need to get to this point but a failure in exec will fail the test
    }
  }

  "Akka Producer and Consumer" should {
    "send string to broker and consume that string back in different consumer groups" in {
      val testMessage = UUID.randomUUID().toString
      val testTopic = UUID.randomUUID().toString
      val groupId_1 = UUID.randomUUID().toString
      val groupId_2 = UUID.randomUUID().toString

      var testStatus1 = false
      var testStatus2 = false

      info("starting akka producertesting")
      val system = ActorSystem("testing")

      val actorCount = 1

      val producer = system.actorOf(Props[KafkaAkkaProducer].withRouter(RoundRobinRouter(actorCount)), "router")

      1 to actorCount foreach { i =>(
        producer ! (testTopic,"192.168.86.10:9092"))
      }

      producer ! testMessage

      val consumer1 = new KafkaConsumer(testTopic,groupId_1,"192.168.86.5:2181")

      def exec1(binaryObject: Array[Byte]) = {
        val message1 = new String(binaryObject)
        info("testMessage 1 = " + testMessage + " and consumed message 1 = " + message1)
        testMessage must_== message1
        consumer1.close()
        testStatus1 = true
      }

      info("KafkaSpec : consumer 1 - is waiting some seconds")
      consumer1.read(exec1)
      info("KafkaSpec : consumer 1 - consumed")

      val consumer2 = new KafkaConsumer(testTopic,groupId_2,"192.168.86.5:2181")

      def exec2(binaryObject: Array[Byte]) = {
        val message2 = new String(binaryObject)
        info("testMessage 2 = " + testMessage + " and consumed message 2 = " + message2)
        testMessage must_== message2
        consumer2.close()
        testStatus2 = true
      }

      info("KafkaSpec : consumer 2 - is waiting some seconds")
      consumer2.read(exec2)
      info("KafkaSpec : consumer 2 - consumed")

      testStatus2 must beTrue // we need to get to this point but a failure in exec will fail the test
    }
  }
}

On the producer side I have started to look more into using Akka. The prototype for this implementation is in the test case above “Akka Producer and Consumer” and broken out here

  "Akka Producer and Consumer" should {
    "send string to broker and consume that string back in different consumer groups" in {
      val testMessage = UUID.randomUUID().toString
      val testTopic = UUID.randomUUID().toString
      val groupId_1 = UUID.randomUUID().toString
      val groupId_2 = UUID.randomUUID().toString

      var testStatus1 = false
      var testStatus2 = false

      info("starting akka producertesting")
      val system = ActorSystem("testing")

      val actorCount = 1

      val producer = system.actorOf(Props[KafkaAkkaProducer].withRouter(RoundRobinRouter(actorCount)), "router")

      1 to actorCount foreach { i =>(
        producer ! (testTopic,"192.168.86.10:9092"))
      }

      producer ! testMessage

      val consumer1 = new KafkaConsumer(testTopic,groupId_1,"192.168.86.5:2181")

      def exec1(binaryObject: Array[Byte]) = {
        val message1 = new String(binaryObject)
        info("testMessage 1 = " + testMessage + " and consumed message 1 = " + message1)
        testMessage must_== message1
        consumer1.close()
        testStatus1 = true
      }

      info("KafkaSpec : consumer 1 - is waiting some seconds")
      consumer1.read(exec1)
      info("KafkaSpec : consumer 1 - consumed")

      val consumer2 = new KafkaConsumer(testTopic,groupId_2,"192.168.86.5:2181")

      def exec2(binaryObject: Array[Byte]) = {
        val message2 = new String(binaryObject)
        info("testMessage 2 = " + testMessage + " and consumed message 2 = " + message2)
        testMessage must_== message2
        consumer2.close()
        testStatus2 = true
      }

      info("KafkaSpec : consumer 2 - is waiting some seconds")
      consumer2.read(exec2)
      info("KafkaSpec : consumer 2 - consumed")

      testStatus2 must beTrue // we need to get to this point but a failure in exec will fail the test
    }
  }
}

What is really nice is you can up that actorCount and really start to run test data to analyze.

/*******************************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
********************************************/

Getting started with Apache Mesos and Apache Aurora using Vagrant

December 16, 2013 2 comments

Apache Mesos is a cluster manager that provides efficient resource isolation and sharing across distributed applications, or frameworks. Think of it as the “kernel” for your data center. Paco Nathan talked about this on one of the All Things Hadoop Podcasts.

Features:

  • Fault-tolerant replicated master using ZooKeeper
  • Scalability to 10,000s of nodes
  • Isolation between tasks with Linux Containers
  • Multi-resource scheduling (memory and CPU aware)
  • Java, Python and C++ APIs for developing new parallel applications
  • Web UI for viewing cluster state

Apache Aurora is a service scheduler that runs on top of Mesos, enabling you to run long-running services that take advantage of Mesos’ scalability, fault-tolerance, and resource isolation. Apache Aurora is currently a part of the Apache Incubator.  The main benefits to a Mesos scheduler like Aurora (and Marathon) is not having to worry about using the Mesos API to take advantage of the grid.  Your application can work the way it does today while Mesos figures out what server(s) to run it on and when to scale that differently from the scheduler.

Features:

  • Deployment and scheduling of jobs
  • The abstraction a “job” to bundle and manage Mesos tasks
  • A rich DSL to define services
  • Health checking
  • Failure domain diversity
  • Instant provisioning

First you need to make sure that you have vagrant and virtual box installed, if you don’t already have these installed then install them please.

1) Install Vagrant http://www.vagrantup.com/
2) Install Virtual Box https://www.virtualbox.org/

That is all you need (assuming you also have git installed). Everything else from here is going to be done from within the virtual machine.


git clone https://github.com/apache/incubator-aurora
cd incubator-aurora
vagrant up

The virtual machines will take some time to spin up so hang tight.

Once the virtual machines are launched you will have your command prompt back and ready to go.

There are 5 vms that are launched: devtools, zookeeper, mesos-master, mesos-slave and aurora-scheduler and they are all configured and networked together (for more info on this check out the Vagrantfile).

Next step is to create an app on the scheduler to provision it to the Mesos cluster that is running.


vagrant ssh aurora-scheduler
vagrant@precise64:~$ cd /vagrant/examples/jobs/
vagrant@precise64:~$aurora create example/www-data/prod/hello hello_world.aurora
 INFO] Creating job hello
 INFO] Response from scheduler: OK (message: 1 new tasks pending for job www-data/prod/hello)
 INFO] Job url: http://precise64:8081/scheduler/www-data/prod/hello

Now go to your browser and pull up http://192.168.33.5:8081/scheduler/www-data/prod/hello and you’ll see your job running

Screen Shot 2013-12-16 at 6.09.42 AM

Basically all of what is happening is in the configuration

hello = Process(
  name = 'hello',
  cmdline = """
    while true; do
      echo hello world
      sleep 10
    done
  """)

task = SequentialTask(
  processes = [hello],
  resources = Resources(cpu = 1.0, ram = 128*MB, disk = 128*MB))

jobs = [Service(
  task = task, cluster = 'example', role = 'www-data', environment = 'prod', name = 'hello')]

It is an exciting time for virtualization and resource scheduling and process provisioning within infrastructures. It is all open source so go dig in and see how it all works for yourself.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

Using Vagrant to get up and running with Apache Kafka

December 7, 2013 Leave a comment

If you haven’t used Vagrant before then hopefully this will not only help you get up and running with Apache Kafka but also get introduced to Vagrant.  Vagrant helps to create and configure lightweight, reproducible, and portable development environments. With Vagrant run a single command — “vagrant up” — and sit back as Vagrant puts together your complete development environment. Say goodbye to the “works on my machine” excuse as Vagrant creates identical development environments for everyone on your team. To learn more about Apache Kafka checkout our website.

To get up and running with Apache Kafka using Vagrant

1) Install Vagrant http://www.vagrantup.com/
2) Install Virtual Box https://www.virtualbox.org/

For the latest release of Apache Kafka 0.8.0 I have added the ability to use Vagrant.

1) git clone https://github.com/stealthly/kafka
2) cd kafka
3) ./sbt update
4) ./sbt package
5) ./sbt assembly-package-dependency
6) vagrant up

Once this is done (it takes a little while for vagrant to launch the four virtual machines)

  • Zookeeper will be running 192.168.50.5
  • Broker 1 on 192.168.50.10
  • Broker 2 on 192.168.50.20
  • Broker 3 on 192.168.50.30

When you are all up and running you will be back at a command prompt.

If you want you can login to the machines using vagrant ssh <machineName> but you don’t need to.

You can access the brokers and zookeeper by their IP

e.g.


bin/kafka-create-topic.sh --zookeeper 192.168.50.5:2181 --replica 3 --partition 1 --topic sandbox

bin/kafka-console-producer.sh --broker-list 192.168.50.10:9092,192.168.50.20:9092,192.168.50.30:9092 --topic sandbox

bin/kafka-console-consumer.sh --zookeeper 192.168.50.5:2181 --topic sandbox --from-beginning

A patch is pending for the 0.8.1 release to have this go back upstream.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/
Categories: Open Source Projects

Technology Decisions Are About Trade Offs and Solving Problems

December 6, 2013 Leave a comment

At some point in the last decade we hit the inflection point where distributed systems, and all their complexities, became the common reality.

Maybe it was the need to change how we scale since CPU clocks are not getting any faster… Maybe it was the Google MapReduce and/or Amazon Dynamo papers… Or maybe it was just the RedSox winning the world series. It doesn’t really matter because we now live in a world where distributed computing is accessible to everyone.

Often the technical arguments around distributed computing are not wrong just rather they do not have proper context. Solutions presented should explain how the technologies/languages solve problems based on the context of their implementation and the trade offs they make.

So much has become solution, solution, solution, without separating and getting down to the problems and then linking problem to solution. Solutions (or technical warnings) come across best when you explain them in context and explain the downside of the solution (so what trade off you accepted for your context).

A few quick examples of this before we continue

 

  • Defending locks as necessary, even superior to a lockless environment. Locks have many well-known and well-documented downsides. They are also used in many systems and have well-documented and well-known benefits. So, context is key here. There are plenty of times when it makes sense to use locks to solve a particular problem. Many people probably have the same problem you have and it is often a complex implementation so they can learn from what you did with and/or without locks that was wrong and what you did that was right.
  • Ripping a language by pointing out everything that is bad about it and saying not to use it. Don’t just talk about problems without providing another solution. Sure, that language might not make sense (for you) when building some type of specific application, but that doesn’t mean it has zero benefits.
  • Promoting how you had a problem with using a technology so instead of realizing that maybe you were doing something wrong you went in and started to question that technology… like sending an email to a mailing list saying that you have “doubts” about the technology when maybe you just don’t understand it.

Just because it doesn’t make sense for you and the context of your domain and that your not willing to accept the trade offs doesn’t make technologies or languages bad or wrong… nor does it make it right.

I love a good informed discussion and debate about technologies that contrast and overlap.  I especially like it when it is not clear which is better for a specific use case. It’s even better when two (or more) technologies each only solve 80% of the problem and do so differently.  This leaves you having to write your own software. And THAT is where the magic happens!

If you can’t talk about the downside and the trade offs then you’re just a fan person waiving a coffee mug around.  That type of behavior really doesn’t benefit anyone.

Context Matters

 

Amarillo Slim said it best “If you can’t spot the sucker within the first half hour at the table, then you are the sucker.”

So, If your’re building a website for your baseball card collection then it doesn’t matter if you use FileMaker Pro, MySQL, Oracle, MongoDB, Cassandra or HBase for your backend. Only your friends and family are going to care if it goes down, loses data or can’t handle load when your entire Facebook and Twitter circles decide to login all at the same time…

Now, lets say you’re building a website (off the top of my head) for signing individuals up for healthcare (hypothetically). Now, if I suggested to write that system in COBOL on a bank of AS400s then I think it would be fair to say that everyone would think that I am crazy and be like “Joe, the sky is blue stop trying to argue it is not”. A COBOL system running on an AS400 is fine if you can find resources willing to work on it and support it with enough capital investment to work around the trade offs.

At the end of the day, computer problems are solved by people and how they apply the technologies as much as the technologies and languages themselves. If you expect tens of thousands or even millions of people all trying to sign up at once (because you have a big opening day launch) then it really doesn’t matter what your system does after that nor what it was written in, what backend it has whether it is cloud based or bare metal or whatever because if they can’t even login or signup then its a bust. The problems nor the trade offs were originally understood so arguing about the technologies/languages being wrong or right is superfluous. The root cause for the failure was a lack of understanding of the problem.

Every technology can be argued to have a use… so let’s stop trashing technologies and be a bit more understanding and focused with our approach. Boil it down to placing technologies (and languages) into their proper category for what they could be good for and what they are not good for. This applies to both cheering a technology/language along and warning people from it. And don’t forget academics and scholarly (both in teaching and research/advancement of the field) solutions. You might do something in school that make sense because it is the best way to learn but that is not actually the solution you would use in other domains.

Focus on the problems and manage the acceptable trade offs to solve those problems

 

So after you get past “is this technically possible” you then have to decide on trade offs and objectively how it affects your context.

If you are not looking at the trade offs then you are doing something wrong. There are always trade offs.

  • What are your real goals? Define them! Be clear with what you are trying to achieve and focus on the problems. Don’t create problems to solve…. Please Please Please, don’t create problems to solve that are not domain driven. And if you do, then don’t promote them as solutions.
  • How much downtime can you afford? Don’t say zero because that is not reality for anyone. Quantify and explain this. Don’t just go and say zero down time rather figure out per second how much money you lose by being down… or what else you might lose for your system being down for how long and quantify and communicate that and why. Apply this not just to the system but each sub-system and possibly every component within each sub-system.
  • How much data is acceptable to lose? If the answer is none then how much availability of the system are you willing to sacrifice? You can have zero loss of data but not without sacrificing availability; however, if availability is more important than how much data you are willing to lose then you have to accept the trade off. For more on this check out Henry’ Robinson’s CAP Confusion: Problems with ‘partition tolerance’ blog post “Partition tolerance means simply developing a coping strategy by choosing which of the other system properties to drop. This is the real lesson of the CAP theorem – if you have a network that may drop messages, then you cannot have both availability and consistency, you must choose one”
  • What are the expectations of the users? Explain this because users provide the domain and the context for what your system does and doesn’t do. What are real requirements vs nice/fancy/flashy/things to have?
  • What is the makeup of your existing team? What is the market for potential new hires? What if you try to pick all the shiny “new awesome” technologies and languages for a solution? What happens when you have to scale the team and you are not able to hire anyone that knows that skill set and no one is interested in adding it to their resume? Is it still shiny and awesome? On the flip side are you using tech that could be antiquated so even though your existing resources are comfortable using the antiquated tech will you have trouble hiring talent moving forward to support it?
  • Do you have existing code and an existing business? Do you have to maintain the code and business with new projects? What can you throw away?  How long do you have to wait before you can throw something away?
  • Are you solving problems you really have or problems you perceive to have?

“Programmers waste enormous amounts of time thinking about, or worrying about, the speed of noncritical parts of their programs, and these attempts at efficiency actually have a strong negative impact when debugging and maintenance are considered. We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%.” – Donald Knuth

Technology Decisions Are About Trade Offs and Solving Problems

 

For me I have my own opinions and my own abilities for solving problems. I don’t always get to work on the use cases I like and sometimes sure, I get use cases that really just completely and totally suck that require me to use technologies and languages that make me vomit in my mouth a little. But, when its all done, if I can still look at myself in the mirror and be proud and feel accomplished with the end result and also that the expectations were exceeded without compromising too much with the trade offs made then solution == awesome.

So with all my rants and thoughts and opinions I leave you to focus on your problems and explain them in context to what you are doing and the trade offs you accepted. Thanks to Camille Fournier, Jay Kreps, Alex Popescu and Steven Gravitz for reviewing this blog post to keep me honest and pull me off the ledge and help with some garbage collection in the process.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

Using Cassandra from Scala

November 16, 2013 5 comments

I have started a new open source project https://github.com/stealthly/scala-cassandra that is a Scala wrapper for CQL  specifically its a wrapper of the DataStax java-driver.

Test cases are a good entry point so lets start here https://github.com/stealthly/scala-cassandra/blob/master/src/test/scala/ScalaCassandraSpec.scala

class ScalaCassandraSpec extends Specification {

    CQL.init()
    CQL.startup("MetaStore")
    Meta.createTable()

    "Meta objects" should {
      "be able to store & retrieve their binary state" in {

      //we use a Thrift object here for portability of the data stored
      val tMeta = new TMeta()

      //setting up some randomness so we can confirm what we are writing is what we get back
      val metaUUID = UUID.randomUUID()
      val dataum = Random.alphanumeric.take(1000000).mkString

      tMeta.setId(metaUUID.toString)
      tMeta.setDatum(dataum)

      Meta.save(tMeta) //Saved to C*

      val someNewTMeta = Meta(metaUUID)
      someNewTMeta.getId() must_== metaUUID.toString

      someNewTMeta.getDatum() must_== dataum
      }
    }
}

The Meta class (contained in the https://github.com/stealthly/scala-cassandra/blob/master/src/main/scala/MetaDAO.scala file) is like a Widget… Whats a Widget? For us lets consider its like “sample code”.

When you want to add something different than a widget create a new thrift file in the `thrift/interface` directory

Then in that directory (lets say you created an IDL called Music.thrift) then run `thrift -gen java Music.thrift` which we did for you already and copy the jar outputted from a `mvn package` to the `lib` folder so you don’t have to worry about it. You can even just keep using the Meta implementaiotn and just shove JSON or XML or whatever you want into it; however, it makes more sense to partition the objects some so you can create wider rows with a collection key in your table along with your partition key. You can also flatten the things by storing a HashMap on the table and retrieving it.

http://stealth.ly

Using Apache Drill for Large Scale, Interactive, Real-Time Analytic Queries

October 29, 2013 Leave a comment

Episode #17 of the podcast is a talk with Jacques Nadeau  available also on iTunes

Apache Drill http://incubator.apache.org/drill/, a modern interactive query engine that runs on top of Hadoop.

Jacques talked about how Apache Drill is a modern query engine that is meant to be a query layer on top of all big data open source systems. Apache Drill is being designed to make the storage engine as plug-able so it could be the interface for any big data storage engine. The first release came out recently to allow developers to understand the data pipeline.

Leveraging an efficient columnar storage format, an optimistic execution engine and a cache-conscious memory layout, Apache Drill is blazing fast. Coordination, query planning, optimization, scheduling, and execution are all distributed throughout nodes in a system to maximize parallelization.

drill_runtime

Perform interactive analysis on all of your data, including nested and schema-less. Drill supports querying against many different schema-less data sources including HBase, Cassandra and MongoDB. Naturally flat records are included as a special case of nested data.

json

Strongly defined tiers and APIs for straightforward integration with a wide array of technologies.

arch

Subscribe to the podcast and listen to what Jacques had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Real-Time Data Pipelines and Analytics with Apache Kafka and Apache Samza

September 17, 2013 Leave a comment

Episode #16 of the podcast is a talk with Jay Kreps  available also on iTunes

Jay talked about the open source work he has done while @ LinkedIn

Including

Most of the conversation was about the Apache Kafka pipeline and the use of Apache Samza for processing it.

Apache Kafka

http://kafka.apache.org/documentation.html#introduction

Kafka is a distributed, partitioned, replicated commit log service. It provides the functionality of a messaging system, but with a unique design.

What does all that mean?

First let’s review some basic messaging terminology:

  • Kafka maintains feeds of messages in categories called topics.
  • We’ll call processes that publish messages to a Kafka topic producers.
  • We’ll call processes that subscribe to topics and process the feed of published messages consumers..
  • Kafka is run as a cluster comprised of one or more servers each of which is called a broker.

So, at a high level, producers send messages over the network to the Kafka cluster which in turn serves them up to consumers like this:

Communication between the clients and the servers is done with a simple, high-performance, language agnosticTCP protocol. We provide a java client for Kafka, but clients are available in many languages.

Topics and Logs

Let’s first dive into the high-level abstraction Kafka provides—the topic.

A topic is a category or feed name to which messages are published. For each topic, the Kafka cluster maintains a partitioned log that looks like this:

Each partition is an ordered, immutable sequence of messages that is continually appended to—a commit log. The messages in the partitions are each assigned a sequential id number called the offset that uniquely identifies each message within the partition.

The Kafka cluster retains all published messages—whether or not they have been consumed—for a configurable period of time. For example if the log retention is set to two days, then for the two days after a message is published it is available for consumption, after which it will be discarded to free up space. Kafka’s performance is effectively constant with respect to data size so retaining lots of data is not a problem.

In fact the only metadata retained on a per-consumer basis is the position of the consumer in in the log, called the “offset”. This offset is controlled by the consumer: normally a consumer will advance its offset linearly as it reads messages, but in fact the position is controlled by the consumer and it can consume messages in any order it likes. For example a consumer can reset to an older offset to reprocess.

This combination of features means that Kafka consumers are very cheap—they can come and go without much impact on the cluster or on other consumers. For example, you can use our command line tools to “tail” the contents of any topic without changing what is consumed by any existing consumers.

Apache Samza

http://samza.incubator.apache.org/

Apache Samza is a distributed stream processing framework. It uses Apache Kafka for messaging, and Apache Hadoop YARN to provide fault tolerance, processor isolation, security, and resource management.

  • Simple API: Unlike most low-level messaging system APIs, Samza provides a very simple call-back based “process message” API that should be familiar to anyone that’s used Map/Reduce.
  • Managed state: Samza manages snapshotting and restoration of a stream processor’s state. Samza will restore a stream processor’s state to a snapshot consistent with the processor’s last read messages when the processor is restarted.
  • Fault tolerance: Samza will work with YARN to restart your stream processor if there is a machine or processor failure.
  • Durability: Samza uses Kafka to guarantee that messages will be processed in the order they were written to a partition, and that no messages will ever be lost.
  • Scalability: Samza is partitioned and distributed at every level. Kafka provides ordered, partitioned, re-playable, fault-tolerant streams. YARN provides a distributed environment for Samza containers to run in.
  • Pluggable: Though Samza works out of the box with Kafka and YARN, Samza provides a pluggable API that lets you run Samza with other messaging systems and execution environments.
  • Processor isolation: Samza works with Apache YARN, which supports processor security through Hadoop’s security model, and resource isolation through Linux CGroups.

Check out Hello Samza to try Samza. Read the Background page to learn more about Samza.

Subscribe to the podcast and listen to what Jay had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Big Data, Open Source and Analytics

August 26, 2013 Leave a comment

Episode #15 of the podcast is a talk with Stefan Groschupf  available also on iTunes

Stefan is the CEO of Datameer and talked about how the company started and where it is now. Founded in 2009 by some of the original contributors to Apache Hadoop, Datameer has grown to a global team, advancing big data analytics. After several implementations of Hadoop analytics solutions at Global 500 companies, the founders were determined to build the next generation analytics application to solve the new use cases created by the explosion of structured and unstructured data. Datameer is the single application for big data analytics by combining data integration, data transformation and data visualization. Customers love us and we work to make Datameer even better each day.

Datameer provides the most complete solution to analyze structured and unstructured data. Not limited by a pre-built schema, the point and click functions means your analytics are only limited by your imagination. Even the most complex nested joins of a large number of datasets can be performed using an interactive dialog. Mix and match analytics and data transformations in unlimited number of data processing pipelines. Leave the raw data untouched.

Datameer turbocharges time series analytics by correlating multiple sets of complex, disparate data. Resulting analytics are endless including correlation of credit card transactions with card holder authorizations, network traffic data, marketing interaction data and many more. The end game is a clear window into the operations of your business, giving you the actionable insights you need to make business decisions.

some alt text

Data is the raw materials of insight and the more data you have, the deeper and broader the possible insights. Not just traditional, transaction data but all types of data so that you can get a complete view of your customers, better understand business processes and improve business performance.

Datameer ignores the limitations of ETL and static schemas to empower business users to integrate data from any source into Hadoop. Pre-built data connector wizards for all common structured and unstructured data sources means that data integration is an easy, three step process of where, what and when.

App Market Infographics

Now you never have to waste precious time by starting from scratch. Anyone can simply browse the Analytics App Market, download an app, connect to data, and get instant results. But why stop there? Every application is completely open so you can customize it, extend it, or even mash it up with other applications to get the insights you need.

Built by data scientists, analysts, or subject matter experts, analytic apps range from horizontal use cases like email and social sentiment analysis to vertical or even product-specific applications like advanced Salesforce.com sales-cycle analysis.

Check out the Datameer app market.

Subscribe to the podcast and listen to what Stefan had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Categories: Hadoop, Podcast

SQL Compatibility in Hadoop with Hive

August 15, 2013 Leave a comment

Episode #14 of the podcast is a talk with Alan Gates available also on iTunes

The Stinger initiative is a collection of development threads in the Hive community that will deliver 100X performance improvements as well as SQL compatibility.

Fast Interactive Query
An immediate aim of 100x performance increase for Hive is more ambitious than any other effort.
SQL Compatibility
Based on industry standard SQL, the Stinger Initiative improves HiveQL to deliver SQL compatibility.

Apache Hive is the de facto standard for SQL-in-Hadoop today with more enterprises relying on this open source project than any alternative. As Hadoop gains in popularity, enterprise requirements for Hive to become more real time or interactive have evolved… and the Hive community has responded.

He spoke in detail about the Stinger initiative, who is contributing to it, why they decided to improve upon Hive and not create a new system and more.

He talked about how Microsoft is contributing in the open source community to improve upon Hive.

Hadoop is so much more than just SQL, one of the wonderful things about Big Data is the power it brings for users to bring different processing models such as realtime streaming with Storm, Graph processing with Giraph and ETL with Pig and all different things to-do beyond just this SQL compatibility.

Alan also talked about YARN and Tez and the benefits of the Stinger initiative to other Hadoop ecosystem tools too.

Subscribe to the podcast and listen to what Alan had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Categories: Hadoop, Hive, Podcast

Apache Zookeeper, Distributed Systems, Open Source and more with Camille Fournier

August 13, 2013 Leave a comment

Episode #13 of the podcast is a talk with Camille Fournier Available also on iTunes

Apache Zookeeper is a centralized service for maintaining configuration information, naming, providing distributed synchronization, and providing group services. All of these kinds of services are used in some form or another by distributed applications. Each time they are implemented there is a lot of work that goes into fixing the bugs and race conditions that are inevitable. Because of the difficulty of implementing these kinds of services, applications initially usually skimp on them ,which make them brittle in the presence of change and difficult to manage.  Even when done correctly, different implementations of these services lead to management complexity when the applications are deployed.

Camille talked about discovery services, distributed locking as well as some tips to developing against and operating Zookeeper in production including how to build a Global, Highly Available Service Discovery Infrastructure with ZooKeeper which she also wrote about on her blog http://whilefalse.blogspot.com/2012/12/building-global-highly-available.html.

Camille gave some great insights about how to apply Open Source community practices to an organization’s SDLC to foster a better culture for better products and services where all developers need to own more parts of their software  (like it is in Open Source projects). #devops #qaops #userops

Subscribe to the podcast and listen to what Camille had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Categories: Podcast, Zookeeper

Apache BigTop and how packaging infrastructure binds the Hadoop ecosystem together

August 12, 2013 Leave a comment

Episode #12 of the podcast is a talk with Mark Grover and Roman Shaposhnik  Available also on iTunes

Apache Bigtop is a project for the development of packaging and tests of the Apache Hadoop ecosystem.

The primary goal of Bigtop is to build a community around the packaging and interoperability testing of Hadoop-related projects. This includes testing at various levels (packaging, platform, runtime, upgrade, etc…) developed by a community with a focus on the system as a whole, rather than individual projects.

BigTop makes it easier to deploy Hadoop Ecosystem projects including:

  • Apache Zookeeper

  • Apache Flume

  • Apache HBase

  • Apache Pig

  • Apache Hive

  • Apache Sqoop

  • Apache Oozie

  • Apache Whirr

  • Apache Mahout

  • Apache Solr (SolrCloud)

  • Apache Crunch (incubating)

  • Apache HCatalog

  • Apache Giraph

  • LinkedIn DataFu

  • Cloudera Hue

The list of supported Linux platforms has expanded to include:

  • CentOS/RHEL 5 and 6

  • Fedora 17 and 18

  • SuSE Linux Enterprise 11

  • OpenSUSE 12.2

  • Ubuntu LTS Lucid (10.04) and Precise (12.04)

  • Ubuntu Quantal (12.10)

Subscribe to the podcast and listen to what Mark and Roman had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Hadoop as a Service cloud platform with the Mortar Framework and Pig

August 9, 2013 Leave a comment

Episode #11 of the podcast is a talk with K Young.  Available also on iTunes

Mortar is the fastest and easiest way to work with Pig and Python on Hadoop in the Cloud.

Mortar’s platform is for everything from joining and cleansing large data sets to machine learning and building recommender systems.

Mortar makes it easy for developers and data scientists to do powerful work with Hadoop. The main advantages of Mortar are:

  • Zero Setup Time: Mortar takes only minutes to set up (or no time at all on the web), and you can start running Pig jobs immediately. No need for painful installation or configuration.
  • Powerful Tooling: Mortar provides a rich suite of tools to aid in Pig development, including the ability to Illustrate a script before running it, and an extremely fast and free local development mode.
  • Elastic Clusters: We spin up Hadoop clusters as you need them, so you don’t have to predict your needs in advance, and you don’t pay for machines you don’t use.
  • Solid Support: Whether the issue is in your script or in Hadoop, we’ll help you figure out a solution.

We talked about the Open Source Mortar Framework and their new Open Source tool for visualizing data while writing Pig scripts called Watchtower

The Mortar Blog has a great video demo on Watchtower.

There are no two ways around it, Hadoop development iterations are slow. Traditional programmers have always had the benefit of re-compiling their app, running it, and seeing the results within seconds. They have near instant validation that what they’re building is actually working. When you’re working with Hadoop, dealing with gigabytes of data, your development iteration time is more like hours.

Subscribe to the podcast and listen to what K Young had to say.  Available also on iTunes

/*********************************
Joe Stein
Founder, Principal Consultant
Big Data Open Source Security LLC
http://www.stealth.ly
Twitter: @allthingshadoop
**********************************/

Categories: Hadoop, Pig, Podcast
Follow

Get every new post delivered to your Inbox.

Join 49 other followers