This article is by Stephen Mouring Jr, appearing courtesy of Scott Leberknight. This is part two of a two part blog series on how to read/write Apache Hive data from MapReduce jos. Part one (Writing Hive Tables from MapReduce) is here. So just as sometimes you need to write data to Hive with a custom MapReduce job, sometimes you need to read that data back from Hive with a custom MapReduce job. As covered in part one, Hive is a layer that sits on HDFS and imposes a standard convention on the structure of the files so it can interpret them as columns and rows. Reading data out of Hive is just a matter of parsing the files correctly. Recall that files processed by MapReduce (and by extension, Hive) are output as key value pairs. Hive ignores the keys (read as a BytesWritable with a value of null) and reads/writes the values as Text objects. The value of the Text object for each row is the concatenation of all the column values delimited by the delimiter of the table (which Hive defaults to the "char 1" ASCII character). Seems like a simple problem, so my first thought was to just using String.split() in the map() method of the MapReduce job. String SEPARATOR_FIELD = new String(new char[] {1}); String[] rowColumns = new String (rowTextObject.getBytes()).split(SEPARATOR_FIELD); In theory this should have worked perfectly, but unfortunately I have found that String.split() actually consumes repeated delimiters. This is a problem if any of the values in the row are blank, since split() will shift the positions of your columns and you will be unable to match up what values belong with which columns. An alternative would be to create a String from the Text object and iterate through it using indexOf(). This approach however requires extra object creation and depending on the scale of your MapReduce job and the size of your rows, may slow you down needlessly. So an alternative is to use the Text object's find() method. String SEPARATOR_FIELD = new String(new char[] {1}); String[] rowColumns = new String[NUMBER_OF_COLUMNS_IN_YOUR_HIVE_TABLE]; int start = 0; int end = 0; for (int i = 0; i < rowColumns.length; ++i) { end = rowTextObject.find(SEPARATOR_FIELD, start); if (end == -1) { end = rowString.getLength(); } rowColumns[i] = new String(rowTextObject.getBytes(), start, end-start); start = end + 1; } This will parse out each value into the appropriately index of the rowColumns array. Blank values will also be handled correctly and result in blank strings being inserted into the rowColumns array.

That’s a pretty long title, I agree. When working on my implementation of RFC2324, also known as the HyperText Coffee Pot Control Protocol, I’ve been struggling with something that you will struggle with as well in your ASP.NET Web API’s: supporting additional HTTP methods like HEAD, PATCH or PROPFIND. ASP.NET Web API has no issue with those, but when hosting them on IIS you’ll find yourself in Yellow-screen-of-death heaven. The reason why IIS blocks these methods (or fails to route them to ASP.NET) is because it may happen that your IIS installation has some configuration leftovers from another API: WebDAV. WebDAV allows you to work with a virtual filesystem (and others) using a HTTP API. IIS of course supports this (because flagship product “SharePoint” uses it, probably) and gets in the way of your API. Bottom line of the story: if you need those methods or want to provide your own HTTP methods, here’s the bit of configuration to add to your Web.config file: Here’s what each part does: Under modules, the WebDAVModule is being removed. Just to make sure that it’s not going to get in our way ever again. The security/requestFiltering element I’ve added only applies if you want to define your own HTTP methods. So unless you need the XYZ method I’ve defined here, don’t add it to your config. Under handlers, I’m removing the default handlers that route into ASP.NET. Then, I’m adding them again. The important part? The "verb attribute. You can provide a list of comma-separated methods that you want to route into ASP.NET. Again, I’ve added my XYZ methodbut you probably don’t need it. This will work on any IIS server as well as on Windows Azure Websites. It will make your API… happy.

The ExecutorService feature had come with Java 5 and is under the java.util.concurrent package. It extends the Executor interface and provides a thread pool functionality to execute asynchronous short tasks. Java Executor Service Types is suggested to look over basic ExecutorService implementation. Also ThreadPoolExecutor is a very useful implementation of ExecutorService ınterface. It extends AbstractExecutorService providing default implementations of ExecutorService execution methods. It provides improved performance when executing large numbers of asynchronous tasks and maintains basic statistics, such as the number of completed tasks. How to develop and monitor Thread Pool Services by using Spring is also suggested to investigate how to develop and monitor Thread Pool Services. So far, we have just talked Undistributed Executor Service implementation. Let us also investigate Distributed Executor Service. Hazelcast Distributed Executor Service feature is a distributed implementation of java.util.concurrent.ExecutorService. It allows to execute business logic in cluster. There are four alternative ways to realize it : 1) The logic can be executed on a specific cluster member which is chosen. 2) The logic can be executed on the member owning the key which is chosen. 3) The logic can be executed on the member Hazelcast will pick. 4) The logic can be executed on all or subset of the cluster members. This article shows how to develop Distributed Executor Service via Hazelcast and Spring. Used Technologies : JDK 1.7.0_09 Spring 3.1.3 Hazelcast 2.4 Maven 3.0.4 STEP 1 : CREATE MAVEN PROJECT A maven project is created as below. (It can be created by using Maven or IDE Plug-in). STEP 2 : LIBRARIES Firstly, Spring dependencies are added to Maven’ s pom.xml 3.1.3.RELEASE UTF-8 org.springframework spring-core ${spring.version} org.springframework spring-context ${spring.version} com.hazelcast hazelcast-all 2.4 log4j log4j 1.2.16 maven-compiler-plugin(Maven Plugin) is used to compile the project with JDK 1.7 org.apache.maven.plugins maven-compiler-plugin 3.0 1.7 1.7 maven-shade-plugin(Maven Plugin) can be used to create runnable-jar org.apache.maven.plugins maven-shade-plugin 2.0 package shade com.onlinetechvision.exe.Application META-INF/spring.handlers META-INF/spring.schemas STEP 3 : CREATE Customer BEAN A new Customer bean is created. This bean will be distributed between two node in OTV cluster. In the following sample, all defined properties(id, name and surname)’ types are String and standart java.io.Serializable interface has been implemented for serializing. If custom or third-party object types are used, com.hazelcast.nio.DataSerializable interface can be implemented for better serialization performance. package com.onlinetechvision.customer; import java.io.Serializable; /** * Customer Bean. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class Customer implements Serializable { private static final long serialVersionUID = 1856862670651243395L; private String id; private String name; private String surname; public String getId() { return id; } public void setId(String id) { this.id = id; } public String getName() { return name; } public void setName(String name) { this.name = name; } public String getSurname() { return surname; } public void setSurname(String surname) { this.surname = surname; } @Override public int hashCode() { final int prime = 31; int result = 1; result = prime * result + ((id == null) ? 0 : id.hashCode()); result = prime * result + ((name == null) ? 0 : name.hashCode()); result = prime * result + ((surname == null) ? 0 : surname.hashCode()); return result; } @Override public boolean equals(Object obj) { if (this == obj) return true; if (obj == null) return false; if (getClass() != obj.getClass()) return false; Customer other = (Customer) obj; if (id == null) { if (other.id != null) return false; } else if (!id.equals(other.id)) return false; if (name == null) { if (other.name != null) return false; } else if (!name.equals(other.name)) return false; if (surname == null) { if (other.surname != null) return false; } else if (!surname.equals(other.surname)) return false; return true; } @Override public String toString() { return "Customer [id=" + id + ", name=" + name + ", surname=" + surname + "]"; } } STEP 4 : CREATE ICacheService INTERFACE A new ICacheService Interface is created for service layer to expose cache functionality. package com.onlinetechvision.cache.srv; import com.hazelcast.core.IMap; import com.onlinetechvision.customer.Customer; /** * A new ICacheService Interface is created for service layer to expose cache functionality. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public interface ICacheService { /** * Adds Customer entries to cache * * @param String key * @param Customer customer * */ void addToCache(String key, Customer customer); /** * Deletes Customer entries from cache * * @param String key * */ void deleteFromCache(String key); /** * Gets Customer cache * * @return IMap Coherence named cache */ IMap getCache(); } STEP 5 : CREATE CacheService IMPLEMENTATION CacheService is implementation of ICacheService Interface. package com.onlinetechvision.cache.srv; import com.hazelcast.core.IMap; import com.onlinetechvision.customer.Customer; import com.onlinetechvision.test.listener.CustomerEntryListener; /** * CacheService Class is implementation of ICacheService Interface. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class CacheService implements ICacheService { private IMap customerMap; /** * Constructor of CacheService * * @param IMap customerMap * */ @SuppressWarnings("unchecked") public CacheService(IMap customerMap) { setCustomerMap(customerMap); getCustomerMap().addEntryListener(new CustomerEntryListener(), true); } /** * Adds Customer entries to cache * * @param String key * @param Customer customer * */ @Override public void addToCache(String key, Customer customer) { getCustomerMap().put(key, customer); } /** * Deletes Customer entries from cache * * @param String key * */ @Override public void deleteFromCache(String key) { getCustomerMap().remove(key); } /** * Gets Customer cache * * @return IMap Coherence named cache */ @Override public IMap getCache() { return getCustomerMap(); } public IMap getCustomerMap() { return customerMap; } public void setCustomerMap(IMap customerMap) { this.customerMap = customerMap; } } STEP 6 : CREATE IDistributedExecutorService INTERFACE A new IDistributedExecutorService Interface is created for service layer to expose distributed execution functionality. package com.onlinetechvision.executor.srv; import java.util.Collection; import java.util.Set; import java.util.concurrent.Callable; import java.util.concurrent.ExecutionException; import com.hazelcast.core.Member; /** * A new IDistributedExecutorService Interface is created for service layer to expose distributed execution functionality. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public interface IDistributedExecutorService { /** * Executes the callable object on stated member * * @param Callable callable * @param Member member * @throws InterruptedException * @throws ExecutionException * */ String executeOnStatedMember(Callable callable, Member member) throws InterruptedException, ExecutionException; /** * Executes the callable object on member owning the key * * @param Callable callable * @param Object key * @throws InterruptedException * @throws ExecutionException * */ String executeOnTheMemberOwningTheKey(Callable callable, Object key) throws InterruptedException, ExecutionException; /** * Executes the callable object on any member * * @param Callable callable * @throws InterruptedException * @throws ExecutionException * */ String executeOnAnyMember(Callable callable) throws InterruptedException, ExecutionException; /** * Executes the callable object on all members * * @param Callable callable * @param Set all members * @throws InterruptedException * @throws ExecutionException * */ Collection executeOnMembers(Callable callable, Set members) throws InterruptedException, ExecutionException; } STEP 7 : CREATE DistributedExecutorService IMPLEMENTATION DistributedExecutorService is implementation of IDistributedExecutorService Interface. package com.onlinetechvision.executor.srv; import java.util.Collection; import java.util.Set; import java.util.concurrent.Callable; import java.util.concurrent.ExecutionException; import java.util.concurrent.ExecutorService; import java.util.concurrent.Future; import java.util.concurrent.FutureTask; import org.apache.log4j.Logger; import com.hazelcast.core.DistributedTask; import com.hazelcast.core.Member; import com.hazelcast.core.MultiTask; /** * DistributedExecutorService Class is implementation of IDistributedExecutorService Interface. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class DistributedExecutorService implements IDistributedExecutorService { private static final Logger logger = Logger.getLogger(DistributedExecutorService.class); private ExecutorService hazelcastDistributedExecutorService; /** * Executes the callable object on stated member * * @param Callable callable * @param Member member * @throws InterruptedException * @throws ExecutionException * */ @SuppressWarnings("unchecked") public String executeOnStatedMember(Callable callable, Member member) throws InterruptedException, ExecutionException { logger.debug("Method executeOnStatedMember is called..."); ExecutorService executorService = getHazelcastDistributedExecutorService(); FutureTask task = (FutureTask) executorService.submit( new DistributedTask(callable, member)); String result = task.get(); logger.debug("Result of method executeOnStatedMember is : " + result); return result; } /** * Executes the callable object on member owning the key * * @param Callable callable * @param Object key * @throws InterruptedException * @throws ExecutionException * */ @SuppressWarnings("unchecked") public String executeOnTheMemberOwningTheKey(Callable callable, Object key) throws InterruptedException, ExecutionException { logger.debug("Method executeOnTheMemberOwningTheKey is called..."); ExecutorService executorService = getHazelcastDistributedExecutorService(); FutureTask task = (FutureTask) executorService.submit(new DistributedTask(callable, key)); String result = task.get(); logger.debug("Result of method executeOnTheMemberOwningTheKey is : " + result); return result; } /** * Executes the callable object on any member * * @param Callable callable * @throws InterruptedException * @throws ExecutionException * */ public String executeOnAnyMember(Callable callable) throws InterruptedException, ExecutionException { logger.debug("Method executeOnAnyMember is called..."); ExecutorService executorService = getHazelcastDistributedExecutorService(); Future task = executorService.submit(callable); String result = task.get(); logger.debug("Result of method executeOnAnyMember is : " + result); return result; } /** * Executes the callable object on all members * * @param Callable callable * @param Set all members * @throws InterruptedException * @throws ExecutionException * */ public Collection executeOnMembers(Callable callable, Set members) throws ExecutionException, InterruptedException { logger.debug("Method executeOnMembers is called..."); MultiTask task = new MultiTask(callable, members); ExecutorService executorService = getHazelcastDistributedExecutorService(); executorService.execute(task); Collection results = task.get(); logger.debug("Result of method executeOnMembers is : " + results.toString()); return results; } public ExecutorService getHazelcastDistributedExecutorService() { return hazelcastDistributedExecutorService; } public void setHazelcastDistributedExecutorService(ExecutorService hazelcastDistributedExecutorService) { this.hazelcastDistributedExecutorService = hazelcastDistributedExecutorService; } } STEP 8 : CREATE TestCallable CLASS TestCallable Class shows business logic to be executed. TestCallable task for first member of the cluster : package com.onlinetechvision.task; import java.io.Serializable; import java.util.concurrent.Callable; /** * TestCallable Class shows business logic to be executed. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class TestCallable implements Callable, Serializable{ private static final long serialVersionUID = -1839169907337151877L; /** * Computes a result, or throws an exception if unable to do so. * * @return String computed result * @throws Exception if unable to compute a result */ public String call() throws Exception { return "First Member' s TestCallable Task is called..."; } } TestCallable task for second member of the cluster : package com.onlinetechvision.task; import java.io.Serializable; import java.util.concurrent.Callable; /** * TestCallable Class shows business logic to be executed. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class TestCallable implements Callable, Serializable{ private static final long serialVersionUID = -1839169907337151877L; /** * Computes a result, or throws an exception if unable to do so. * * @return String computed result * @throws Exception if unable to compute a result */ public String call() throws Exception { return "Second Member' s TestCallable Task is called..."; } } STEP 9 : CREATE AnotherAvailableMemberNotFoundException CLASS AnotherAvailableMemberNotFoundException is thrown when another available member is not found. To avoid this exception, first node should be started before the second node. package com.onlinetechvision.exception; /** * AnotherAvailableMemberNotFoundException is thrown when another available member is not found. * To avoid this exception, first node should be started before the second node. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class AnotherAvailableMemberNotFoundException extends Exception { private static final long serialVersionUID = -3954360266393077645L; /** * Constructor of AnotherAvailableMemberNotFoundException * * @param String Exception message * */ public AnotherAvailableMemberNotFoundException(String message) { super(message); } } STEP 10 : CREATE CustomerEntryListener CLASS CustomerEntryListener Class listens entry changes on named cache object. package com.onlinetechvision.test.listener; import com.hazelcast.core.EntryEvent; import com.hazelcast.core.EntryListener; /** * CustomerEntryListener Class listens entry changes on named cache object. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ @SuppressWarnings("rawtypes") public class CustomerEntryListener implements EntryListener { /** * Invoked when an entry is added. * * @param EntryEvent * */ public void entryAdded(EntryEvent ee) { System.out.println("EntryAdded... Member : " + ee.getMember() + ", Key : "+ee.getKey()+", OldValue : "+ee.getOldValue()+", NewValue : "+ee.getValue()); } /** * Invoked when an entry is removed. * * @param EntryEvent * */ public void entryRemoved(EntryEvent ee) { System.out.println("EntryRemoved... Member : " + ee.getMember() + ", Key : "+ee.getKey()+", OldValue : "+ee.getOldValue()+", NewValue : "+ee.getValue()); } /** * Invoked when an entry is evicted. * * @param EntryEvent * */ public void entryEvicted(EntryEvent ee) { } /** * Invoked when an entry is updated. * * @param EntryEvent * */ public void entryUpdated(EntryEvent ee) { } } STEP 11 : CREATE Starter CLASS Starter Class loads Customers to cache and executes distributed tasks. Starter Class of first member of the cluster : package com.onlinetechvision.exe; import com.onlinetechvision.cache.srv.ICacheService; import com.onlinetechvision.customer.Customer; /** * Starter Class loads Customers to cache and executes distributed tasks. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class Starter { private ICacheService cacheService; /** * Loads cache and executes the tasks * */ public void start() { loadCacheForFirstMember(); } /** * Loads Customers to cache * */ public void loadCacheForFirstMember() { Customer firstCustomer = new Customer(); firstCustomer.setId("1"); firstCustomer.setName("Jodie"); firstCustomer.setSurname("Foster"); Customer secondCustomer = new Customer(); secondCustomer.setId("2"); secondCustomer.setName("Kate"); secondCustomer.setSurname("Winslet"); getCacheService().addToCache(firstCustomer.getId(), firstCustomer); getCacheService().addToCache(secondCustomer.getId(), secondCustomer); } public ICacheService getCacheService() { return cacheService; } public void setCacheService(ICacheService cacheService) { this.cacheService = cacheService; } } Starter Class of second member of the cluster : package com.onlinetechvision.exe; import java.util.Set; import java.util.concurrent.ExecutionException; import com.hazelcast.core.Hazelcast; import com.hazelcast.core.HazelcastInstance; import com.hazelcast.core.Member; import com.onlinetechvision.cache.srv.ICacheService; import com.onlinetechvision.customer.Customer; import com.onlinetechvision.exception.AnotherAvailableMemberNotFoundException; import com.onlinetechvision.executor.srv.IDistributedExecutorService; import com.onlinetechvision.task.TestCallable; /** * Starter Class loads Customers to cache and executes distributed tasks. * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class Starter { private String hazelcastInstanceName; private Hazelcast hazelcast; private IDistributedExecutorService distributedExecutorService; private ICacheService cacheService; /** * Loads cache and executes the tasks * */ public void start() { loadCache(); executeTasks(); } /** * Loads Customers to cache * */ public void loadCache() { Customer firstCustomer = new Customer(); firstCustomer.setId("3"); firstCustomer.setName("Bruce"); firstCustomer.setSurname("Willis"); Customer secondCustomer = new Customer(); secondCustomer.setId("4"); secondCustomer.setName("Colin"); secondCustomer.setSurname("Farrell"); getCacheService().addToCache(firstCustomer.getId(), firstCustomer); getCacheService().addToCache(secondCustomer.getId(), secondCustomer); } /** * Executes Tasks * */ public void executeTasks() { try { getDistributedExecutorService().executeOnStatedMember(new TestCallable(), getAnotherMember()); getDistributedExecutorService().executeOnTheMemberOwningTheKey(new TestCallable(), "3"); getDistributedExecutorService().executeOnAnyMember(new TestCallable()); getDistributedExecutorService().executeOnMembers(new TestCallable(), getAllMembers()); } catch (InterruptedException | ExecutionException | AnotherAvailableMemberNotFoundException e) { e.printStackTrace(); } } /** * Gets cluster members * * @return Set Set of Cluster Members * */ private Set getAllMembers() { Set members = getHazelcastLocalInstance().getCluster().getMembers(); return members; } /** * Gets an another member of cluster * * @return Member Another Member of Cluster * @throws AnotherAvailableMemberNotFoundException An Another Available Member can not found exception */ private Member getAnotherMember() throws AnotherAvailableMemberNotFoundException { Set members = getAllMembers(); for(Member member : members) { if(!member.localMember()) { return member; } } throw new AnotherAvailableMemberNotFoundException("No Other Available Member on the cluster. Please be aware that all members are active on the cluster"); } /** * Gets Hazelcast local instance * * @return HazelcastInstance Hazelcast local instance */ @SuppressWarnings("static-access") private HazelcastInstance getHazelcastLocalInstance() { HazelcastInstance instance = getHazelcast().getHazelcastInstanceByName(getHazelcastInstanceName()); return instance; } public String getHazelcastInstanceName() { return hazelcastInstanceName; } public void setHazelcastInstanceName(String hazelcastInstanceName) { this.hazelcastInstanceName = hazelcastInstanceName; } public Hazelcast getHazelcast() { return hazelcast; } public void setHazelcast(Hazelcast hazelcast) { this.hazelcast = hazelcast; } public IDistributedExecutorService getDistributedExecutorService() { return distributedExecutorService; } public void setDistributedExecutorService(IDistributedExecutorService distributedExecutorService) { this.distributedExecutorService = distributedExecutorService; } public ICacheService getCacheService() { return cacheService; } public void setCacheService(ICacheService cacheService) { this.cacheService = cacheService; } } STEP 12 : CREATE hazelcast-config.properties FILE hazelcast-config.properties file shows the properties of cluster members. First member properties : hz.instance.name = OTVInstance1 hz.group.name = dev hz.group.password = dev hz.management.center.enabled = true hz.management.center.url = http://localhost:8080/mancenter hz.network.port = 5701 hz.network.port.auto.increment = false hz.tcp.ip.enabled = true hz.members = 192.168.1.32 hz.executor.service.core.pool.size = 2 hz.executor.service.max.pool.size = 30 hz.executor.service.keep.alive.seconds = 30 hz.map.backup.count=2 hz.map.max.size=0 hz.map.eviction.percentage=30 hz.map.read.backup.data=true hz.map.cache.value=true hz.map.eviction.policy=NONE hz.map.merge.policy=hz.ADD_NEW_ENTRY Second member properties : hz.instance.name = OTVInstance2 hz.group.name = dev hz.group.password = dev hz.management.center.enabled = true hz.management.center.url = http://localhost:8080/mancenter hz.network.port = 5702 hz.network.port.auto.increment = false hz.tcp.ip.enabled = true hz.members = 192.168.1.32 hz.executor.service.core.pool.size = 2 hz.executor.service.max.pool.size = 30 hz.executor.service.keep.alive.seconds = 30 hz.map.backup.count=2 hz.map.max.size=0 hz.map.eviction.percentage=30 hz.map.read.backup.data=true hz.map.cache.value=true hz.map.eviction.policy=NONE hz.map.merge.policy=hz.ADD_NEW_ENTRY STEP 13 : CREATE applicationContext-hazelcast.xml Spring Hazelcast Configuration file, applicationContext-hazelcast.xml, is created and Hazelcast Distributed Executor Service and Hazelcast Instance are configured. ${hz.instance.name} ${hz.members} STEP 14 : CREATE applicationContext.xml Spring Configuration file, applicationContext.xml, is created. classpath:/hazelcast-config.properties STEP 15 : CREATE Application CLASS Application Class is created to run the application. ackage com.onlinetechvision.exe; import org.springframework.context.ApplicationContext; import org.springframework.context.support.ClassPathXmlApplicationContext; /** * Application class starts the application * * @author onlinetechvision.com * @since 27 Nov 2012 * @version 1.0.0 * */ public class Application { /** * Starts the application * * @param String[] args * */ public static void main(String[] args) { ApplicationContext context = new ClassPathXmlApplicationContext("applicationContext.xml"); Starter starter = (Starter) context.getBean("starter"); starter.start(); } } STEP 16 : BUILD PROJECT After OTV_Spring_Hazelcast_DistributedExecution Project is built, OTV_Spring_Hazelcast_DistributedExecution-0.0.1-SNAPSHOT.jar will be created. Important Note : The Members of the cluster have got different configuration for Coherence so the project should be built separately for each member. STEP 17 : INTEGRATION with HAZELCAST MANAGEMENT CENTER Hazelcast Management Center enables to monitor and manage nodes in the cluster. Entity and backup counts which are owned by customerMap, can be seen via Map Memory Data Table. We have distributed 4 entries via customerMap as shown below : Sample keys and values can be seen via Map Browser : Added First Entry : Added Third Entry : hazelcastDistributedExecutorService details can be seen via Executors tab. We have executed 3 task on first member and 2 tasks on second member as shown below : STEP 18 : RUN PROJECT BY STARTING THE CLUSTER’ s MEMBER After created OTV_Spring_Hazelcast_DistributedExecution-0.0.1-SNAPSHOT.jar file is run at the cluster’ s members, the following console output logs will be shown : First member console output : Kas 25, 2012 4:07:20 PM com.hazelcast.impl.AddressPicker INFO: Interfaces is disabled, trying to pick one address from TCP-IP config addresses: [x.y.z.t] Kas 25, 2012 4:07:20 PM com.hazelcast.impl.AddressPicker INFO: Prefer IPv4 stack is true. Kas 25, 2012 4:07:20 PM com.hazelcast.impl.AddressPicker INFO: Picked Address[x.y.z.t]:5701, using socket ServerSocket[addr=/0:0:0:0:0:0:0:0,localport=5701], bind any local is true Kas 25, 2012 4:07:21 PM com.hazelcast.system INFO: [x.y.z.t]:5701 [dev] Hazelcast Community Edition 2.4 (20121017) starting at Address[x.y.z.t]:5701 Kas 25, 2012 4:07:21 PM com.hazelcast.system INFO: [x.y.z.t]:5701 [dev] Copyright (C) 2008-2012 Hazelcast.com Kas 25, 2012 4:07:21 PM com.hazelcast.impl.LifecycleServiceImpl INFO: [x.y.z.t]:5701 [dev] Address[x.y.z.t]:5701 is STARTING Kas 25, 2012 4:07:24 PM com.hazelcast.impl.TcpIpJoiner INFO: [x.y.z.t]:5701 [dev] --A new cluster is created and First Member joins the cluster. Members [1] { Member [x.y.z.t]:5701 this } Kas 25, 2012 4:07:24 PM com.hazelcast.impl.MulticastJoiner INFO: [x.y.z.t]:5701 [dev] Members [1] { Member [x.y.z.t]:5701 this } ... -- First member adds two new entries to the cache... EntryAdded... Member : Member [x.y.z.t]:5701 this, Key : 1, OldValue : null, NewValue : Customer [id=1, name=Jodie, surname=Foster] EntryAdded... Member : Member [x.y.z.t]:5701 this, Key : 2, OldValue : null, NewValue : Customer [id=2, name=Kate, surname=Winslet] ... --Second Member joins the cluster. Members [2] { Member [x.y.z.t]:5701 this Member [x.y.z.t]:5702 } ... -- Second member adds two new entries to the cache... EntryAdded... Member : Member [x.y.z.t]:5702, Key : 4, OldValue : null, NewValue : Customer [id=4, name=Colin, surname=Farrell] EntryAdded... Member : Member [x.y.z.t]:5702, Key : 3, OldValue : null, NewValue : Customer [id=3, name=Bruce, surname=Willis] Second member console output : Kas 25, 2012 4:07:48 PM com.hazelcast.impl.AddressPicker INFO: Interfaces is disabled, trying to pick one address from TCP-IP config addresses: [x.y.z.t] Kas 25, 2012 4:07:48 PM com.hazelcast.impl.AddressPicker INFO: Prefer IPv4 stack is true. Kas 25, 2012 4:07:48 PM com.hazelcast.impl.AddressPicker INFO: Picked Address[x.y.z.t]:5702, using socket ServerSocket[addr=/0:0:0:0:0:0:0:0,localport=5702], bind any local is true Kas 25, 2012 4:07:49 PM com.hazelcast.system INFO: [x.y.z.t]:5702 [dev] Hazelcast Community Edition 2.4 (20121017) starting at Address[x.y.z.t]:5702 Kas 25, 2012 4:07:49 PM com.hazelcast.system INFO: [x.y.z.t]:5702 [dev] Copyright (C) 2008-2012 Hazelcast.com Kas 25, 2012 4:07:49 PM com.hazelcast.impl.LifecycleServiceImpl INFO: [x.y.z.t]:5702 [dev] Address[x.y.z.t]:5702 is STARTING Kas 25, 2012 4:07:49 PM com.hazelcast.impl.Node INFO: [x.y.z.t]:5702 [dev] ** setting master address to Address[x.y.z.t]:5701 Kas 25, 2012 4:07:49 PM com.hazelcast.impl.MulticastJoiner INFO: [x.y.z.t]:5702 [dev] Connecting to master node: Address[x.y.z.t]:5701 Kas 25, 2012 4:07:49 PM com.hazelcast.nio.ConnectionManager INFO: [x.y.z.t]:5702 [dev] 55715 accepted socket connection from /x.y.z.t:5701 Kas 25, 2012 4:07:55 PM com.hazelcast.cluster.ClusterManager INFO: [x.y.z.t]:5702 [dev] --Second Member joins the cluster. Members [2] { Member [x.y.z.t]:5701 Member [x.y.z.t]:5702 this } Kas 25, 2012 4:07:56 PM com.hazelcast.impl.LifecycleServiceImpl INFO: [x.y.z.t]:5702 [dev] Address[x.y.z.t]:5702 is STARTED -- Second member adds two new entries to the cache... EntryAdded... Member : Member [x.y.z.t]:5702 this, Key : 3, OldValue : null, NewValue : Customer [id=3, name=Bruce, surname=Willis] EntryAdded... Member : Member [x.y.z.t]:5702 this, Key : 4, OldValue : null, NewValue : Customer [id=4, name=Colin, surname=Farrell] 25.11.2012 16:07:56 DEBUG (DistributedExecutorService.java:42) - Method executeOnStatedMember is called... 25.11.2012 16:07:56 DEBUG (DistributedExecutorService.java:46) - Result of method executeOnStatedMember is : First Member' s TestCallable Task is called... 25.11.2012 16:07:56 DEBUG (DistributedExecutorService.java:61) - Method executeOnTheMemberOwningTheKey is called... 25.11.2012 16:07:56 DEBUG (DistributedExecutorService.java:65) - Result of method executeOnTheMemberOwningTheKey is : First Member' s TestCallable Task is called... 25.11.2012 16:07:56 DEBUG (DistributedExecutorService.java:78) - Method executeOnAnyMember is called... 25.11.2012 16:07:57 DEBUG (DistributedExecutorService.java:82) - Result of method executeOnAnyMember is : Second Member' s TestCallable Task is called... 25.11.2012 16:07:57 DEBUG (DistributedExecutorService.java:96) - Method executeOnMembers is called... 25.11.2012 16:07:57 DEBUG (DistributedExecutorService.java:101) - Result of method executeOnMembers is : [First Member' s TestCallable Task is called..., Second Member' s TestCallable Task is called...] STEP 19 : DOWNLOAD https://github.com/erenavsarogullari/OTV_Spring_Hazelcast_DistributedExecution REFERENCES : Java ExecutorService Interface Hazelcast Distributed Executor Service

A frequent issue I come across writing integration applications with Mule is deciding how to communicate back and forth between my front end application, typically a web or mobile application, and a flow hosted on Mule. I could use web services and do something like annotate a component with JAX-RS and expose this out over HTTP. This is potentially overkill, particularly if I only want to host a few methods, the methods are asynchronous or I don’t want to deal with the overhead of HTTP. It also could be a lot of extra effort if the only consumers of the API, at least initially, are internal facing applications. Another choice is to use “synchronous” JMS with temporary reply queues. While Mule makes this easy to do, particularly with MuleClient, I now have to deal with the overhead of spinning up a JMS infrastructure. I could also be limited to Java only clients, depending on which JMS broker I choose. The latter is particularly signifcant, as Java probably isn’t the technology of choice on the web or mobile layer. ØMQ for RPC ØMQ, or ZeroMQ, is a networking library designed from the ground up to ease integration between distributed applications. In addition to supporting a variety of messaging patterns, which are enumerated in the extremely well written guide, the library is written in platform agnostic C with wrappers for different languages like Java, Python and Ruby. These features make it a good candidate to solve the challenges I introduced above, particularly since a community contributed module for ØMQ was released recently. Let’s consider a simple service that accepts a request for a range of stock quotes and returns the results and see how we can host this service with Mule and expose it out with the ØMQ Module. Data Serialization with Protocol Buffers Data is transported back and forth over ØMQ as byte arrays. We, as such, need to decide on a way to serialize our stock quote request and responses “on the wire.” Before we do that, however, let’s take a look at the Java canonical data model we’re using on the client and server side. The following Gists show the important bits of the StockQuote and StockQuoteResponse classes. public class StockQuote implements Serializable { String symbol; Date date; Double open; Double high; Double low; Double close; Long volume; Double adjustedClose; public class StockQuoteRequest implements Serializable { String symbol; Date startDate; Date endDate; public interface StockDataService { public List getQuote(StockQuoteRequest request); } We could use Java serialization to get the objects into byte arrays. Ignoring the other deficiencies of default Java serialization, the main drawback is that it limits our clients to one’s running on a JVM. XML or JSON provide better alternatives, but for the purposes of this example we’ll assume we want a more compact representation of the data (this isn’t totally unrealistic, stock quote data can be extremely time sensitive and we probably want to minimize serialization and deserialization overhead.) Protocol Buffers provide a good middle ground and also boast a Mule Module to provide the necessary transformers we need to move back and forth from the byte array representations. Let’s define two .proto files to define the wire format and generate the intermediary stubs for serialization. package com.acmesoft.zeromq; option java_package = "com.acmesoft.stock.model.serialization.protobuf"; option optimize_for = SPEED;package com.acmesoft.zeromq; option java_package = "com.acmesoft.stock.model.serialization.protobuf"; option optimize_for = SPEED; option java_multiple_files = true; message StockQuoteResponseBuffer { repeated StockQuoteBuffer result = 1; } message StockQuoteBuffer { required string symbol = 1; required int64 date = 2; required double open = 3; required double high = 4; required double low = 5; required double close = 6; required int64 volume = 7; required double adjustedClose = 8; } option java_multiple_files = true; message StockQuoteRequestBuffer { required string symbol = 1; required int64 start = 2; required int64 end = 3; } You typically would use the “protoc” compiler to generate the Java stubs. This is tedious, however, so we’ll instead modify the pom.xml of our project to compile the protoc files during the compile goals: com.google.protobuf.tools maven-protoc-plugin /usr/local/bin/protoc compile testCompile Since we already have a domain model we’ll add some helper classes to simplify the serialization tasks on the client side. public byte[] toProtocolBufferAsBytes() { return StockQuoteRequestBuffer.newBuilder() .setSymbol(symbol) .setStart(startDate.getTime()) .setEnd(endDate.getTime()).build().toByteArray(); } public static StockQuoteRequest fromProtocolBuffer(StockQuoteRequestBuffer buffer) { StockQuoteRequest request = new StockQuoteRequest(); request.setSymbol(buffer.getSymbol()); request.setStartDate(new Date(buffer.getStart())); request.setEndDate(new Date(buffer.getEnd())); return request; } public static StockQuoteResponseBuffer toProtocolBuffer(List quotes) { StockQuoteResponseBuffer.Builder responseBuilder = StockQuoteResponseBuffer.newBuilder(); for (StockQuote quote : quotes) { responseBuilder.addResult(StockQuoteBuffer.newBuilder() .setAdjustedClose(quote.getAdjustedClose()) .setClose(quote.getClose()) .setDate(quote.getDate().getTime()) .setHigh(quote.getHigh()) .setLow(quote.getLow()) .setOpen(quote.getOpen()) .setSymbol(quote.getSymbol()) .setVolume(quote.getVolume()).build()); } return responseBuilder.build(); } public static List listOfStockQuotesFromBytes(byte[] bytes) { List buffer; try { buffer = StockQuoteResponseBuffer.parseFrom(bytes).getResultList(); } catch (InvalidProtocolBufferException e) { throw new SerializationException(e); } List quotes = new ArrayList(); for (StockQuoteBuffer stockQuoteBuffer : buffer) { StockQuote stockQuote = new StockQuote(); stockQuote.setClose(stockQuoteBuffer.getClose()); stockQuote.setDate(new Date(stockQuoteBuffer.getDate())); stockQuote.setHigh(stockQuoteBuffer.getHigh()); stockQuote.setOpen(stockQuoteBuffer.getOpen()); stockQuote.setSymbol(stockQuoteBuffer.getSymbol()); stockQuote.setVolume(stockQuoteBuffer.getVolume()); stockQuote.setAdjustedClose(stockQuoteBuffer.getAdjustedClose()); stockQuote.setLow(stockQuoteBuffer.getLow()); quotes.add(stockQuote); } return quotes; } Configuring StockDataService Now that we have a canonical data model and a wire format defined we’re ready to wire up a Mule flow to expose the service out. Note that for this to work you need to have jzmq installed locally on your system. The following dependency needs to be added to your pom.xml once its installed: org.zeromq zmq 2.2.0 /usr/local/lib/zmq.jar system Where systemPath is the location of the zmq.jar on your filesystem. Once that’s out of the way we can configure the flow, as illustrated below: The ZeroMQ inbound-endpoint will be bound to TCP port 9090 with a request-response exchange pattern. The deserialize MP in the protobuf module will deserialize the byte array to the generated StockQuoteRequestBuffer class. From there we’ll use MEL to invoke the helper method on StockQuoteRequest to transform the intermediary class to the domain model. The List of StockQuotes returned from StockDataService will be transformed by the MEL expression using the “toProtocolBuffer” helper method on the domain model. The Protocol Buffer Module is then smart enough to implicitly transform the intermediary object to a byte array for the response. Consuming the Service from the Client Side Now that the server is ready we can turn our attention to the client side code to invoke the remote service. Let’s take a look at how this works: StockQuoteRequest stockQuoteRequest = new StockQuoteRequest(); stockQuoteRequest.setSymbol("FB"); stockQuoteRequest.setStartDate(new Date( new Date().getTime() - (86400000 * 7))); stockQuoteRequest.setEndDate(new Date()); ZMQ.Socket zmqSocket = zmqContext.socket(ZMQ.REQ); zmqSocket.setReceiveTimeOut(RECEIVE_TIMEOUT); zmqSocket.connect("tcp://localhost:9090"); zmqSocket.send(stockQuoteRequest.toProtocolBufferAsBytes(), 0); List quotes = StockQuote.listOfStockQuotesFromBytes(zmqSocket.recv(0)); We start off by defining the StockQuoteRequest object to give us all the quotes for Facebook stock from the last week. We can then open up a ZMQ socket, set the timeout, connect to the ZMQ socket on the remote Mule instance and send the byte representation of the StockQuoteRequest to it. zmqSocket.recv is then used to receive the bytes back from Mule. From here we can use the listOfStockQuotesFromBytes helper method we wrote above to convert the Protocol Buffer representation to a List of StockQuotes. Despite the fair bit of plumbing we did above, this is a pretty concise bit of client side code to invoke the remote service. Conclusion This blog post only touched on the features of ØMQ and the ØMQ Mule Module. In addition to request-reply, other exchange-patterns are supported, like one-way, push and pull. This effectively gives you the benefits of a reliable, asynchronous messaging layer without a centralized infrastructure. I hope to cover this in a later post. Protocol buffers also seem like a natural fit as a wire format for ØMQ. protobuffers echo ØMQ’s principals of being lightweight, fast and platform agnostic. These are also, not coincidently, principals Mule shares as an integration framework. The project for this example is available on GitHub.

We've talked previously about the Windows Azure command-line tools, and have used them in a few posts such as Brian's Migrating Drupal to a Windows Azure VM. While the tools are generally useful for tons of stuff, one of the things that's been painful to do with the command-line is export the settings for a VM, and then recreate the VM from those settings. You might be wondering why you'd want to export a VM and then recreate it. For me, cost is the first thing that comes to mind. It costs more to keep a VM running than it does to just keep the disk in storage. So if I had something in a VM that I'm only using a few hours a day, I'd delete the VM when I'm not using it and recreate it when I need it again. Another potential reason is that you want to create a copy of the disk so that you can create a duplicate virtual machine. The export process used to be pretty arcane stuff; using the azure vm show command with a --json parameter and piping the output to file. Then hacking the .json file to fix it up so it could be used with the azure vm create-from command. It was bad. It was so bad, the developers added a new export command to create the .json file for you. Here's the basic process: Create a VM VM creation has been covered multiple ways already; you're either going to use the portal or command line tools, and you're either going to select an image from the library or upload a VHD. In my case, I used the following command: azure vm create larryubuntu CANONICAL__Canonical-Ubuntu-12-04-amd64-server-20120528.1.3-en-us-30GB.vhd larry NotaRe This command creates a new VM in the East US data center, enables SSH on port 22 and then stores a disk image for this VM in a blob. You can see the new disk image in blob storage by running: azure vm disk list The results should return something like: info: Executing command vm disk list + Fetching disk images data: Name OS data: ---------------------------------------- ------- data: larryubuntu-larryubuntu-0-20121019170709 Linux info: vm disk list command OK That's the actual disk image that is mounted by the VM. Export and Delete the VM Alright, I've done my work and it's the weekend. I need to export the VM settings so I can recreate it on Monday, then delete the VM so I won't get charged for the next 48 hours of not working. To export the settings for the VM, I use the following command: azure vm export larryubuntu c:\stuff\vminfo.json This tells Windows Azure to find the VM named larryubuntu and export its settings to c:\stuff\vminfo.json. The .json file will contain something like this: { "RoleName":"larryubuntu", "RoleType":"PersistentVMRole", "ConfigurationSets": [ { "ConfigurationSetType":"NetworkConfiguration", "InputEndpoints": [ { "LocalPort":"22", "Name":"ssh", "Port":"22", "Protocol":"tcp", "Vip":"168.62.177.227" } ], "SubnetNames":[] } ], "DataVirtualHardDisks":[], "OSVirtualHardDisk": { "HostCaching":"ReadWrite", "DiskName":"larryubuntu-larryubuntu-0-20121024155441", "OS":"Linux" }, "RoleSize":"Small" } If you're like me, you'll immediately start thinking "Hrmmm, I wonder if I can mess around with things like RoleSize." And yes, you can. If you wanted to bump this up to medium, you'd just change that parameter to medium. If you want to play around more with the various settings, it looks like the schema is maintained at https://github.com/WindowsAzure/azure-sdk-for-node/blob/master/lib/services/serviceManagement/models/roleschema.json. Once I've got the file, I can safely delete the VM by using the following command. azure vm delete larryubuntu It spins a bit and then no more VM. Recreate the VM Ugh, Monday. Time to go back to work, and I need my VM back up and running. So I run the following command: azure vm create-from larryubuntu c:\stuff\vminfo.json --location "East US" It takes only a minute or two to spin up the VM and it's ready for work. That's it - fast, simple, and far easier than the old process of generating the .json settings file. Note that I haven't played around much with the various settings described in the schema for the json file that I linked above. If you find anything useful or interesting that can be accomplished by hacking around with the .json, leave a comment about it.

For the past few months, I’ve been coaching a “Microsoft Student Partner” (who has a great blog on Kinect for Windows by the way!) on Windows Azure. One of the questions he recently had was around PartitionKey and RowKey in Windows Azure Table Storage. What are these for? Do I have to specify them manually? Let’s explain… Windows Azure storage partitions All Windows Azure storage abstractions (Blob, Table, Queue) are built upon the same stack (whitepaper here). While there’s much more to tell about it, the reason why it scales is because of its partitioning logic. Whenever you store something on Windows Azure storage, it is located on some partition in the system. Partitions are used for scale out in the system. Imagine that there’s only 3 physical machines that are used for storing data in Windows Azure storage: Based on the size and load of a partition, partitions are fanned out across these machines. Whenever a partition gets a high load or grows in size, the Windows Azure storage management can kick in and move a partition to another machine: By doing this, Windows Azure can ensure a high throughput as well as its storage guarantees. If a partition gets busy, it’s moved to a server which can support the higher load. If it gets large, it’s moved to a location where there’s enough disk space available. Partitions are different for every storage mechanism: In blob storage, each blob is in a separate partition. This means that every blob can get the maximal throughput guaranteed by the system. In queues, every queue is a separate partition. In tables, it’s different: you decide how data is co-located in the system. PartitionKey in Table Storage In Table Storage, you have to decide on the PartitionKey yourself. In essence, you are responsible for the throughput you’ll get on your system. If you put every entity in the same partition (by using the same partition key), you’ll be limited to the size of the storage machines for the amount of storage you can use. Plus, you’ll be constraining the maximal throughput as there’s lots of entities in the same partition. Should you set the PartitionKey to the same value for every entity stored? No. You’ll end up with scaling issues at some point. Should you set the PartitionKey to a unique value for every entity stored? No. You can do this and every entity stored will end up in its own partition, but you’ll find that querying your data becomes more difficult. And that’s where our next concept kicks in… RowKey in Table Storage A RowKey in Table Storage is a very simple thing: it’s your “primary key” within a partition. PartitionKey + RowKey form the composite unique identifier for an entity. Within one PartitionKey, you can only have unique RowKeys. If you use multiple partitions, the same RowKey can be reused in every partition. So in essence, a RowKey is just the identifier of an entity within a partition. PartitionKey and RowKey and performance Before building your code, it’s a good idea to think about both properties. Don’t just assign them a guid or a random string as it does matter for performance. The fastest way of querying? Specifying both PartitionKey and RowKey. By doing this, table storage will immediately know which partition to query and can simply do an ID lookup on RowKey within that partition. Less fast but still fast enough will be querying by specifying PartitionKey: table storage will know which partition to query. Less fast: querying on only RowKey. Doing this will give table storage no pointer on which partition to search in, resulting in a query that possibly spans multiple partitions, possibly multiple storage nodes as well. Wihtin a partition, searching on RowKey is still pretty fast as it’s a unique index. Slow: searching on other properties (again, spans multiple partitions and properties). Note that Windows Azure storage may decide to group partitions in so-called "Range partitions" - see http://msdn.microsoft.com/en-us/library/windowsazure/hh508997.aspx. In order to improve query performance, think about your PartitionKey and RowKey upfront, as they are the fast way into your datasets. Deciding on PartitionKey and RowKey Here’s an exercise: say you want to store customers, orders and orderlines. What will you choose as the PartitionKey (PK) / RowKey (RK)? Let’s use three tables: Customer, Order and Orderline. An ideal setup may be this one, depending on how you want to query everything: Customer (PK: sales region, RK: customer id) – it enables fast searches on region and on customer id Order (PK: customer id, RK; order id) – it allows me to quickly fetch all orders for a specific customer (as they are colocated in one partition), it still allows fast querying on a specific order id as well) Orderline (PK: order id, RK: order line id) – allows fast querying on both order id as well as order line id. Of course, depending on the system you are building, the following may be a better setup: Customer (PK: customer id, RK: display name) – it enables fast searches on customer id and display name Order (PK: customer id, RK; order id) – it allows me to quickly fetch all orders for a specific customer (as they are colocated in one partition), it still allows fast querying on a specific order id as well) Orderline (PK: order id, RK: item id) – allows fast querying on both order id as well as the item bought, of course given that one order can only contain one order line for a specific item (PK + RK should be unique) You see? Choose them wisely, depending on your queries. And maybe an important sidenote: don’t be afraid of denormalizing your data and storing data twice in a different format, supporting more query variations. There’s one additional “index” That’s right! People have been asking Microsoft for a secondary index. And it’s already there… The table name itself! Take our customer – order – orderline sample again… Having a Customer table containing all customers may be interesting to search within that data. But having an Orders table containing every order for every customer may not be the ideal solution. Maybe you want to create an order table per customer? Doing that, you can easily query the order id (it’s the table name) and within the order table, you can have more detail in PK and RK. And there's one more: your account name. Split data over multiple storage accounts and you have yet another "partition". Conclusion In conclusion? Choose PartitionKey and RowKey wisely. The more meaningful to your application or business domain, the faster querying will be and the more efficient table storage will work in the long run.

EasyNetQ, my super simple .NET API for RabbitMQ, now (from version 0.7.2.34) supports RabbitMQ clusters without any need to deploy a load balancer. Simply list the nodes of the cluster in the connection string ... var bus = RabbitHutch.CreateBus("host=ubuntu:5672,ubuntu:5673"); In this example I have set up a cluster on a single machine, 'ubuntu', with node 1 on port 5672 and node 2 on port 5673. When the CreateBus statement executes, EasyNetQ will attempt to connect to the first host listed (ubuntu:5672). If it fails to connect it will attempt to connect to the second host listed (ubuntu:5673). If neither node is available it will sit in a re-try loop attempting to connect to both servers every five seconds. It logs all this activity to the registered IEasyNetQLogger. You might see something like this if the first node was unavailable: DEBUG: Trying to connect ERROR: Failed to connect to Broker: 'ubuntu', Port: 5672 VHost: '/'. ExceptionMessage: 'None of the specified endpoints were reachable' DEBUG: OnConnected event fired INFO: Connected to RabbitMQ. Broker: 'ubuntu', Port: 5674, VHost: '/' If the node that EasyNetQ is connected to fails, EasyNetQ will attempt to connect to the next listed node. Once connected, it will re-declare all the exchanges and queues and re-start all the consumers. Here's an example log record showing one node failing then EasyNetQ connecting to the other node and recreating the subscribers: INFO: Disconnected from RabbitMQ Broker DEBUG: Trying to connect DEBUG: OnConnected event fired DEBUG: Re-creating subscribers INFO: Connected to RabbitMQ. Broker: 'ubuntu', Port: 5674, VHost: '/' You get automatic fail-over out of the box. That’s pretty cool. If you have multiple services using EasyNetQ to connect to a RabbitMQ cluster, they will all initially connect to the first listed node in their respective connection strings. For this reason the EasyNetQ cluster support is not really suitable for load balancing high throughput systems. I would recommend that you use a dedicated hardware or software load balancer instead, if that’s what you want.

Curator's note: This article originally appeared at WindowsAzure.com. To use this feature and other new Windows Azure capabilities, sign up for the free preview. Just as you can quickly create and deploy a web application created from the gallery, you can also deploy a website created on a workstation with traditional developer tools from Microsoft or other companies. Table of Contents Deployment Options How to: Create a Website Using the Management Portal How to: Create a Website from the Gallery How to: Delete a Website Next Steps Deployment Options Windows Azure supports deploying websites from remote computers using WebDeploy, FTP, GIT or TFS. Many development tools provide integrated support for publication using one or more of these methods and may only require that you provide the necessary credentials, site URL and hostname or URL for your chosen deployment method. Credentials and deployment URLs for all enabled deployment methods are stored in the website's publish profile, a file which can be downloaded in the Windows Azure (Preview) Management Portal from the Quick Start page or the quick glance section of the Dashboard page. If you prefer to deploy your website with a separate client application, high quality open source GIT and FTP clients are available for download on the Internet for this purpose. How to: Create a Website Using the Management Portal Follow these steps to create a website in Windows Azure. Login to the Windows Azure (Preview) Management Portal. Click the Create New icon on the bottom left of the Management Portal. Click the Web Site icon, click the Quick Create icon, enter a value for URL and then click the check mark next to create web site on the bottom right corner of the page. When the website has been created you will see the text Creation of Web Site '[SITENAME]' Completed. Click the name of the website displayed in the list of websites to open the website's Quick Start management page. On the Quick Start page you are provided with options to set up TFS or GIT publishing if you would like to deploy your finished website to Windows Azure using these methods. FTP publishing is set up by default for websites and the FTP Host name is displayed under FTP Hostname on the Quick Start and Dashboard pages. Before publishing with FTP or GIT choose the option to Reset deployment credentials on the Dashboard page. Then specify the new credentials (username and password) to authenticate against the FTP Host or the Git Repository when deploying content to the website. The Configure management page exposes several configurable application settings in the following sections: Framework: Set the version of .NET framework or PHP required by your web application. Diagnostics: Set logging options for gathering diagnostic information for your website in this section. App Settings: Specify name/value pairs that will be loaded by your web application on start up. For .NET sites, these settings will be injected into your .NET configuration AppSettings at runtime, overriding existing settings. For PHP and Node sites these settings will be available as environment variables at runtime. Connection Strings: View connection strings for linked resources. For .NET sites, these connection strings will be injected into your .NET configuration connectionStrings settings at runtime, overriding existing entries where the key equals the linked database name. For PHP and Node sites these settings will be available as environment variables at runtime. Default Documents: Add your web application's default document to this list if it is not already in the list. If your web application contains more than one of the files in the list then make sure your website's default document appears at the top of the list. How to: Create a Website from the Gallery The gallery makes available a wide range of popular web applications developed by Microsoft, third party companies, and open source software initiatives. Web applications created from the gallery do not require installation of any software other than the browser used to connect to the Windows Azure Management Portal. In this tutorial, you'll learn: How to create a new site through the gallery. How to deploy the site through the Windows Azure Portal. You'll build a Word press blog that uses a default template. The following illustration shows the completed application: Note To complete this tutorial, you need a Windows Azure account that has the Windows Azure Web Sites feature enabled. You can create a free trial account and enable preview features in just a couple of minutes. For details, see Create a Windows Azure account and enable preview features. Create a web site in the portal Login to the Windows Azure Management Portal. Click the New icon on the bottom left of the dashboard. Click the Web Site icon, and click From Gallery. Locate and click the WordPress icon in list, and then click Next. On the Configure Your App page, enter or select values for all fields: Enter a URL name of your choice Leave Create a new MySQL database selected in the Database field Select the region closest to you Then click Next. On the Create New Database page, you can specify a name for your new MySQL database or use the default name. Select the region closest to you as the hosting location. Select the box at the bottom of the screen to agree to ClearDB's usage terms for your hosted MySQL database. Then click the check to complete the site creation. After you click Complete Windows Azure will initiate build and deploy operations. While the web site is being built and deployed the status of these operations is displayed at the bottom of the Web Sites page. After all operations are performed, A final status message when the site has been successfully deployed. Launch and manage your WordPress site Click on your new site from the Web Sites page to open the dashboard for the site. On the Dashboard management page, scroll down and click the link on the left under Site Url to open the site’s welcome page. Enter appropriate configuration information required by WordPress and click Install WordPress to finalize configuration and open the web site’s login page. Login to the new WordPress web site by entering the username and password that you specified on the Welcome page. You'll have a new WordPress site that looks similar to the site below. How to: Delete a Website Websites are deleted using the Delete icon in the Windows Azure Management Portal. The Delete icon is available in the Windows Azure Portal when you click Web Sites to list all of your websites and at the bottom of each of the website management pages. Next Steps For more information about Websites, see the following: Walkthrough: Troubleshooting a Website on Windows Azure

Hadoop MapReduce is a YARN-based system for parallel processing of large data sets. In this article, learn to quickly start writing the simplest MapReduce job.

In Java, you can easily implement some business logic in Plain Old Java Object (POJO) classes, and then able to run them in a fancy server or framework without much hassle. There many server/frameworks, such as JBossAS, Spring or Camel etc, that would allow you to deploy POJO without even hardcoding to their API. Obviously you would get advance features if you willing to couple to their API specifics, but even if you do, you can keep these to minimal by encapsulating your own POJO and their API in a wrapper. By writing and designing your own application as simple POJO as possible, you will have the most flexible ways in choose a framework or server to deploy and run your application. One effective way to write your business logic in these environments is to use Service component. In this article I will share few things I learned in writing Services. What is a Service? The word Service is overly used today, and it could mean many things to different people. When I say Service, my definition is a software component that has minimal of life-cycles such as init, start, stop, and destroy. You may not need all these stages of life-cycles in every service you write, but you can simply ignore ones that don't apply. When writing large application that intended for long running such as a server component, definining these life-cycles and ensure they are excuted in proper order is crucial! I will be walking you through a Java demo project that I have prepared. It's very basic and it should run as stand-alone. The only dependency it has is the SLF4J logger. If you don't know how to use logger, then simply replace them with System.out.println. However I would strongly encourage you to learn how to use logger effectively during application development though. Also if you want to try out the Spring related demos, then obviously you would need their jars as well. Writing basic POJO service You can quickly define a contract of a Service with life-cycles as below in an interface. package servicedemo; public interface Service { void init(); void start(); void stop(); void destroy(); boolean isInited(); boolean isStarted(); } Developers are free to do what they want in their Service implementation, but you might want to give them an adapter class so that they don't have to re-write same basic logic on each Service. I would provide an abstract service like this: package servicedemo; import java.util.concurrent.atomic.*; import org.slf4j.*; public abstract class AbstractService implements Service { protected Logger logger = LoggerFactory.getLogger(getClass()); protected AtomicBoolean started = new AtomicBoolean(false); protected AtomicBoolean inited = new AtomicBoolean(false); public void init() { if (!inited.get()) { initService(); inited.set(true); logger.debug("{} initialized.", this); } } public void start() { // Init service if it has not done so. if (!inited.get()) { init(); } // Start service now. if (!started.get()) { startService(); started.set(true); logger.debug("{} started.", this); } } public void stop() { if (started.get()) { stopService(); started.set(false); logger.debug("{} stopped.", this); } } public void destroy() { // Stop service if it is still running. if (started.get()) { stop(); } // Destroy service now. if (inited.get()) { destroyService(); inited.set(false); logger.debug("{} destroyed.", this); } } public boolean isStarted() { return started.get(); } public boolean isInited() { return inited.get(); } @Override public String toString() { return getClass().getSimpleName() + "[id=" + System.identityHashCode(this) + "]"; } protected void initService() { } protected void startService() { } protected void stopService() { } protected void destroyService() { } } This abstract class provide the basic of most services needs. It has a logger and states to keep track of the life-cycles. It then delegate new sets of life-cycle methods so subclass can choose to override. Notice that the start() method is checking auto calling init() if it hasn't already done so. Same is done in destroy() method to the stop() method. This is important if we're to use it in a container that only have two stages life-cycles invocation. In this case, we can simply invoke start() and destroy() to match to our service's life-cycles. Some frameworks might go even further and create separate interfaces for each stage of the life-cycles, such as InitableService or StartableService etc. But I think that would be too much in a typical app. In most of the cases, you want something simple, so I like it just one interface. User may choose to ignore methods they don't want, or simply use an adaptor class. Before we end this section, I would throw in a silly Hello world service that can be used in our demo later. package servicedemo; public class HelloService extends AbstractService { public void initService() { logger.info(this + " inited."); } public void startService() { logger.info(this + " started."); } public void stopService() { logger.info(this + " stopped."); } public void destroyService() { logger.info(this + " destroyed."); } } Managing multiple POJO Services with a container Now we have the basic of Service definition defined, your development team may start writing business logic code! Before long, you will have a library of your own services to re-use. To be able group and control these services into an effetive way, we want also provide a container to manage them. The idea is that we typically want to control and manage multiple services with a container as a group in a higher level. Here is a simple implementation for you to get started: package servicedemo; import java.util.*; public class ServiceContainer extends AbstractService { private List services = new ArrayList(); public void setServices(List services) { this.services = services; } public void addService(Service service) { this.services.add(service); } public void initService() { logger.debug("Initializing " + this + " with " + services.size() + " services."); for (Service service : services) { logger.debug("Initializing " + service); service.init(); } logger.info(this + " inited."); } public void startService() { logger.debug("Starting " + this + " with " + services.size() + " services."); for (Service service : services) { logger.debug("Starting " + service); service.start(); } logger.info(this + " started."); } public void stopService() { int size = services.size(); logger.debug("Stopping " + this + " with " + size + " services in reverse order."); for (int i = size - 1; i >= 0; i--) { Service service = services.get(i); logger.debug("Stopping " + service); service.stop(); } logger.info(this + " stopped."); } public void destroyService() { int size = services.size(); logger.debug("Destroying " + this + " with " + size + " services in reverse order."); for (int i = size - 1; i >= 0; i--) { Service service = services.get(i); logger.debug("Destroying " + service); service.destroy(); } logger.info(this + " destroyed."); } } From above code, you will notice few important things: We extends the AbstractService, so a container is a service itself. We would invoke all service's life-cycles before moving to next. No services will start unless all others are inited. We should stop and destroy services in reverse order for most general use cases. The above container implementation is simple and run in synchronized fashion. This mean, you start container, then all services will start in order you added them. Stop should be same but in reverse order. I also hope you would able to see that there is plenty of room for you to improve this container as well. For example, you may add thread pool to control the execution of the services in asynchronized fashion. Running POJO Services Running services with a simple runner program. In the simplest form, we can run our POJO services on our own without any fancy server or frameworks. Java programs start its life from a static main method, so we surely can invoke init and start of our services in there. But we also need to address the stop and destroy life-cycles when user shuts down the program (usually by hitting CTRL+C.) For this, the Java has the java.lang.Runtime#addShutdownHook() facility. You can create a simple stand-alone server to bootstrap Service like this: package servicedemo; import org.slf4j.*; public class ServiceRunner { private static Logger logger = LoggerFactory.getLogger(ServiceRunner.class); public static void main(String[] args) { ServiceRunner main = new ServiceRunner(); main.run(args); } public void run(String[] args) { if (args.length < 1) throw new RuntimeException("Missing service class name as argument."); String serviceClassName = args[0]; try { logger.debug("Creating " + serviceClassName); Class serviceClass = Class.forName(serviceClassName); if (!Service.class.isAssignableFrom(serviceClass)) { throw new RuntimeException("Service class " + serviceClassName + " did not implements " + Service.class.getName()); } Object serviceObject = serviceClass.newInstance(); Service service = (Service)serviceObject; registerShutdownHook(service); logger.debug("Starting service " + service); service.init(); service.start(); logger.info(service + " started."); synchronized(this) { this.wait(); } } catch (Exception e) { throw new RuntimeException("Failed to create and run " + serviceClassName, e); } } private void registerShutdownHook(final Service service) { Runtime.getRuntime().addShutdownHook(new Thread() { public void run() { logger.debug("Stopping service " + service); service.stop(); service.destroy(); logger.info(service + " stopped."); } }); } } With abover runner, you should able to run it with this command: $ java demo.ServiceRunner servicedemo.HelloService Look carefully, and you'll see that you have many options to run multiple services with above runner. Let me highlight couple: Improve above runner directly and make all args for each new service class name, instead of just first element. Or write a MultiLoaderService that will load multiple services you want. You may control argument passing using System Properties. Can you think of other ways to improve this runner? Running services with Spring The Spring framework is an IoC container, and it's well known to be easy to work POJO, and Spring lets you wire your application together. This would be a perfect fit to use in our POJO services. However, with all the features Spring brings, it missed a easy to use, out of box main program to bootstrap spring config xml context files. But with what we built so far, this is actually an easy thing to do. Let's write one of our POJO Service to bootstrap a spring context file. package servicedemo; import org.springframework.context.ConfigurableApplicationContext; import org.springframework.context.support.FileSystemXmlApplicationContext; public class SpringService extends AbstractService { private ConfigurableApplicationContext springContext; public void startService() { String springConfig = System.getProperty("springContext", "spring.xml); springContext = new FileSystemXmlApplicationContext(springConfig); logger.info(this + " started."); } public void stopService() { springContext.close(); logger.info(this + " stopped."); } } With that simple SpringService you can run and load any spring xml file. For example try this: $ java -DspringContext=config/service-demo-spring.xml demo.ServiceRunner servicedemo.SpringService Inside the config/service-demo-spring.xml file, you can easily create our container that hosts one or more service in Spring beans. Notice that I only need to setup init-method and destroy-method once on the serviceContainer bean. You can then add one or more other service such as the helloService as much as you want. They will all be started, managed, and then shutdown when you close the Spring context. Note that Spring context container did not explicitly have the same life-cycles as our services. The Spring context will automatically instanciate all your dependency beans, and then invoke all beans who's init-method is set. All that is done inside the constructor of FileSystemXmlApplicationContext. No explicit init method is called from user. However at the end, during stop of the service, Spring provide the springContext#close() to clean things up. Again, they do not differentiate stop from destroy. Because of this, we must merge our init and start into Spring's init state, and then merge stop and destroy into Spring's close state. Recall our AbstractService#destory will auto invoke stop if it hasn't already done so. So this is trick that we need to understand in order to use Spring effectively. Running services with JEE app server In a corporate env, we usually do not have the freedom to run what we want as a stand-alone program. Instead they usually have some infrustructure and stricter standard technology stack in place already, such as using a JEE application server. In these situation, the most portable way to run POJO services is in a war web application. In a Servlet web application, you can write a class that implements javax.servlet.ServletContextListener and this will provide you the life-cycles hook via contextInitialized and contextDestroyed. In there, you can instanciate your ServiceContainer object and call start and destroy methods accordingly. Here is an example that you can explore: package servicedemo; import java.util.*; import javax.servlet.*; public class ServiceContainerListener implements ServletContextListener { private static Logger logger = LoggerFactory.getLogger(ServiceContainerListener.class); private ServiceContainer serviceContainer; public void contextInitialized(ServletContextEvent sce) { serviceContainer = new ServiceContainer(); List services = createServices(); serviceContainer.setServices(services); serviceContainer.start(); logger.info(serviceContainer + " started in web application."); } public void contextDestroyed(ServletContextEvent sce) { serviceContainer.destroy(); logger.info(serviceContainer + " destroyed in web application."); } private List createServices() { List result = new ArrayList(); // populate services here. return result; } } You may configure above in the WEB-INF/web.xml like this: servicedemo.ServiceContainerListener The demo provided a placeholder that you must add your services in code. But you can easily make that configurable using the web.xml for context parameters. If you were to use Spring inside a Servlet container, you may directly use their org.springframework.web.context.ContextLoaderListener class that does pretty much same as above, except they allow you to specify their xml configuration file using the contextConfigLocation context parameter. That's how a typical Spring MVC based application is configure. Once you have this setup, you can experiment our POJO service just as the Spring xml sample given above to test things out. You should see our service in action by your logger output. PS: Actually what we described here are simply related to Servlet web application, and not JEE specific. So you can use Tomcat server just fine as well. The importance of Service's life-cycles and it's real world usage All the information I presented here are not novelty, nor a killer design pattern. In fact they have been used in many popular open source projects. However, in my past experience at work, folks always manage to make these extremely complicated, and worse case is that they completely disregard the importance of life-cycles when writing services. It's true that not everything you going to write needs to be fitted into a service, but if you find the need, please do pay attention to them, and take good care that they do invoked properly. The last thing you want is to exit JVM without clean up in services that you allocated precious resources for. These would become more disastrous if you allow your application to be dynamically reloaded during deployment without exiting JVM, in which will lead to system resources leakage. The above Service practice has been put into use in the TimeMachine project. In fact, if you look at the timemachine.scheduler.service.SchedulerEngine, it would just be a container of many services running together. And that's how user can extend the scheduler functionalities as well, by writing a Service. You can load these services dynamically by a simple properties file.

DrupalCon Munich is next week, and I am lucky enough to be going. As part of preparing for the conference, I thought it would be worthwhile to see just how easy (or difficult) it would be to migrate an existing Drupal site to Windows Azure Web Sites. So, in this post, I’ll do just that. Fortunately, because Windows Azure Web Sites supports both PHP and MySQL, the migration process is relatively straightforward. And, because Drupal and PHP run on any platform, the process I’ll describe should work for moving Drupal to Windows Azure Web Sites regardless of what platform you are moving from. Of course, Drupal installations can vary widely, so YMMV. I tested the instructions below on relatively small (and simple) Drupal installation running on CentOS 5. (Unfortunately, I won’t be using Drush since it isn’t supported on Windows Azure Websites.) If you are considering moving a large and complex Drupal application, may want to consider moving to Windows Azure Cloud Services (more information about that here: Migrating a Drupal Site from LAMP to Windows Azure). Before getting started, it’s worth noting that Windows Azure Websites lets you run up to 10 Web Sites for free in a multitenant environment. And, you can seamlessly upgrade to private, reserved VM instances as your traffic grows. To sign up, try the Windows Azure 90-day free trial. 1. Create a Windows Azure Web Site and MySQL database There is a step-by-step tutorial on http://www.windowsazure.com that walks you through creating a new website and a MySQL database, so I’ll refer you there to get started: Create a PHP-MySQL Windows Azure web site and deploy using Git. If you intend to use Git to publish your Drupal site, then go ahead and follow the instructions for setting up a Git repository. Make sure to follow the instructions in the Get remote MySQL connection information section as you will need that information later. You can ignore the remainder of the tutorial for the purposes of deploying your Drupal site, but if you are new to Windows Azure Web Sites (and to Git), you might find the additional reading informative. Ok, now you have a new website with a MySQL database, your have your MySQL database connection information, and you have (optionally) created a remote Git repository and made note of the Git deployment instructions. Now you are ready to copy your database to MySQL in Windows Azure Web Sites. 2. Copy database to MySQL in Windows Azure Web Sites I’m sure there is more than one way to copy your Drupal database, but I found the mysqldump tool to be effective and easy to use. To copy from a local machine to Windows Azure Web Sites, here’s the command I used: mysqldump -u local_username --password=local_password drupal | mysql -h remote_host -u remote_username --password=remote_password remote_db_name You will, of course, have to provide the username and password for your existing Drupal database, and you will have to provide the hostname, username, password, and database name for the MySQL database you created in step 1. This information is available in the connection string information that you should have noted in step 1. i.e. You should have a connection string that looks something like this: Database=remote_db_name;Data Source=remote_host;User Id=remote_username;Password=remote_password Depending on the size of your database, the copying process could take several minutes. Now your Drupal database is live in Windows Azure Websites. Before you deploy your Drupal code, you need to modify it so it can connect to the new database. 3. Modify database connection info in settings.php Here, you will again need your new database connection information. Open the /drupal/sites/default/setting.php file in your favorite text editor, and replace the values of ‘database’, ‘username’, ‘password’, and ‘host’ in the $databases array with the correct values for your new database. When you are finished, you should have something similar to this: $databases = array ( 'default' => array ( 'default' => array ( 'database' => 'remote_db_name', 'username' => 'remote_username', 'password' => 'remote_password', 'host' => 'remote_host', 'port' => '', 'driver' => 'mysql', 'prefix' => '', ), ), ); Be sure to save the settings.phpfile, then you are ready to deploy. 4. Deploy Drupal code using Git or FTP The last step is to deploy your code to Windows Azure Web Sites using Git or FTP. If you are using FTP, you can get the FTP hostname and username from you website’s dashboard. Then, use your favorite FTP client to upload your Drupal files to the /site/wwwroot folder of the remote site. If you are using Git, you need to set up a Git repository in Windows Azure Web Sites (steps for this are in the tutorial mentioned earlier). And, you will need Git installed on your local machine. Then, just follow the instructions provided after you created the repository: One note about using Git here: depending on your Git settings, your .gitignore file (a hidden file and a sibling to the .git folder created in your local root directory after you executed git commit), some files in your Drupal application may be ignored. In my case, all the files in the sites directory were ignored. If this happens, you will want to edit the .gitignore file so that these files aren’t ignored and redeploy. After you have deployed Drupal to Windows Azure Web Sites, you can continue to deploy updates via Git or FTP. Related information If you are looking for more information about Windows Azure Web Sites, these posts might be helpful: Windows Azure Websites- A PHP Perspective Windows Azure Websites, Web Roles, and VMs- When to use which- Configuring PHP in Windows Azure Websites with .user.ini Files One last thing you might consider, depending on your site, is using the Windows Azure Integration Module to store and serve your site’s media files.

Autoscaling your webserver tier is typically straightforward. Image your apache server with source code or without, then sync down files from S3 upon spinup. Roll that image into the autoscale configuration and you’re all set. With the database tier though, things can be a bit tricky. The typical configuration we see is to have a single master database where your application writes. But scaling out or horizontally on Amazon EC2 should be as easy as adding more slaves, right? Why not automate that process? Below we’ve set out to answer some of the questions you’re likely to face when setting up slaves against your master. We’ve included instructions on building an AMI that automatically spins up as a slave. Fancy! How can I autoscale my database tier? Build an auto-starting MySQL slave against your master. Configure those to spinup. Amazon’s autoscaling loadbalancer is one option, another is to use a roll-your-own solution, monitoring thresholds on servers, and spinning up or dropping off slaves as necessary. Does an AWS snapshot capture subvolume data or just the SIZE of the attached volume? In fact, if you have an attached EBS volume and you create an new AMI off of that, you will capture the entire root volume, plus your attached volume data. In fact we find this a great way to create an auto-building slave in the cloud. How do I freeze MySQL during AWS snapshot? mysql> flush tables with read lock;mysql> system xfs_freeze -f /data At this point you can use the Amazon web console, ylastic, or ec2-create-image API call to do so from the command line. When the server you are imaging off of above restarts – as it will do by default – it will start with /data partition unfrozen and mysql’s tables unlocked again. Voila! If you’re not using xfs for your /data filesystem, you should be. It’s fast! The xfsprogs docs seem to indicate this may also work with foreign filesystems. Check the docs for details. How do I build an AMI mysql slave that autoconnects to master? Install mysql_serverid script below. Configure mysql to use your /data EBS mount. Set all your my.cnf settings including server_id Configure the instance as a slave in the normal way. When using GRANT to create the ‘rep’ user on master, specify the host with a subnet wildcard. For example ’10.20.%’. That will subsequently allow any 10.20.x.y servers to connect and replicate. Point the slave at the master. When all is running properly, edit the my.cnf file and remove server_id. Don’t restart mysql. Freeze the filesystem as described above. Use the Amazon console, ylastic or API call to create your new image. Test it of course, to make sure it spins up, sets server_id and connects to master. Make a change in the test schema, and verify that it propagates to all slaves. How do I set server_id uniquely? As you hopefully already know, in MySQL replication environment each node requires a unique server_id setting. In my Amazon Machine Images, I want the server to startup and if it doesn’t find the server_id in the /etc/my.cnf file, to add it there, correctly! Is that so much to ask? Here’s what I did. Fire up your editor of choice and drop in this bit of code: #!/bin/shif grep -q “server_id” /etc/my.cnf then : # do nothing – it’s already set else # extract numeric component from hostname – should be internet IP in Amazon environment export server_id=`echo $HOSTNAME | sed ‘s/[^0-9]*//g’` echo “server_id=$server_id” >> /etc/my.cnf # restart mysql /etc/init.d/mysql restart fi Save that snippet at /root/mysql_serverid. Also be sure to make it executable: $ chmod +x /root/mysql_serverid Then just append it to your /etc/rc.local file with an editor or echo: $ echo "/root/mysql_serverid" >> /etc/rc.local Assuming your my.cnf file does *NOT* contain the server_id setting when you re-image, then it’ll set this automagically each time you spinup a new server off of that AMI. Nice! Can you easily slave off of a slave? How? It’s not terribly different from slaving off of a normal master. A. First enable slave updates. The setting is not dynamic, so if you don’t already have it set, you’ll have to restart your slave. log_slave_updates=true B. Get an initial snapshot of your slave data. You can do that the locking way: mysql> flush tables with read lock;mysql> show master status\G; mysql> system mysqldump -A > full_slave_dump.mysql mysql> unlock tables; You may also choose to use Percona’s excellent xtrabackup utility to create hotbackups without locking any tables. We are very lucky to have an open-source tool like this at our disposal. MySQL Enterprise Backup from Oracle Corp can also do this. C. On the slave, seed the database with your dump created above. $ mysql < full_slave_dump.mysql D. Now point your slave to the original slave. mysql> change master to master_user='rep', master_password='rep', master_host='192.168.0.1', master_log_file='server-bin-log.000004', master_log_pos=399;mysql> start slave; mysql> show slave status\G; Slave master is set as an IP address. Is there another way? It’s possible to use hostnames in MySQL replication, however it’s not recommended. Why? Because of the wacky world of DNS. Suffice it to say MySQL has to do a lot of work to resolve those names into IP addresses. A hickup in DNS can interrupt all MySQL services potentially as sessions will fail to authenticate. To avoid this problem do two things: A. Set this parameter in my.cnf skip_name_resolve = true Remove entries in mysql.user table where hostname is not an IP address. Those entries will be invalid for authentication after setting the above parameter. Doesn’t RDS take care of all of this for me? RDS is Amazon’s Relational Database Service which is built on MySQL. Amazon’s RDS solution presents MySQL as a service which brings certain benefits to administrators and startups: Simpler administration. Nuts and bolts are handled for you. Push-button replication. No more struggling with the nuances and issues of MySQL’s replication management. Simplicity of administration of course has it’s downsides. Depending on your environment, these may or may not be dealbreakers. No access to the slow query log. This is huge. The single best tool for troubleshooting slow database response is this log file. Queries are a large part of keeping a relational database server healthy and happy, and without this facility, you are severely limited. Locked in downtime window When you signup for RDS, you must define a thirty minute maintenance window. This is a weekly window during which your instance *COULD* be unavailable. When you host yourself, you may not require as much downtime at all, especially if you’re using master-master mysql and zero-downtime configuration. Can’t use Percona Server to host your MySQL data. You won’t be able to do this in RDS. Percona server is a high performance distribution of MySQL which typically rolls in serious performance tweaks and updates before they make it to community addition. Well worth the effort to consider it. No access to filesystem, server metrics & command line. Again for troubleshooting problems, these are crucial. Gathering data about what’s really happening on the server is how you begin to diagnose and troubleshoot a server stall or pileup. You are beholden to Amazon’s support services if things go awry. That’s because you won’t have access to the raw iron to diagnose and troubleshoot things yourself. Want to call in an outside consultant to help you debug or troubleshoot? You’ll have your hands tied without access to the underlying server. You can’t replicate to a non-RDS database. Have your own datacenter connected to Amazon via VPC? Want to replication to a cloud server? RDS won’t fit the bill. You’ll have to roll your own – as we’ve described above. And if you want to replicate to an alternate cloud provider, again RDS won’t work for you. Related posts: Deploying MySQL on Amazon EC2 – 8 Best Practices Review: Host Your Web Site In The Cloud, Amazon Web Services Made Easy 5 Ways to Boost MySQL Scalability Top MySQL DBA interview questions (Part 2) MySQL Cluster In The Cloud – Managers Guide

Learn all about working with MondoDB multimaster.

After receiving a lot of good feedback and comment on my last blog on MongoDb, I was encouraged to do another deep dive on another popular document oriented db; Couchbase. I have been a long-time fan CouchDb and has wrote a blog on it many years ago. After it merges with Membase, I am very excited to take a deep look into it again. Couchbase is the merge of two popular NOSQL technologies: Membase, which provides persistence, replication, sharding to the high performance memcached technology CouchDB, which pioneers the document oriented model based on JSON Like other NOSQL technologies, both Membase and CouchDB are built from the ground up on a highly distributed architecture, with data shard across machines in a cluster. Built around the Memcached protocol, Membase provides an easy migration to existing Memcached users who want to add persistence, sharding and fault resilience on their familiar Memcached model. On the other hand, CouchDB provides first class support for storing JSON documents as well as a simple RESTful API to access them. Underneath, CouchDB also has a highly tuned storage engine that is optimized for both update transaction as well as query processing. Taking the best of both technologies, Membase is well-positioned in the NOSQL marketplace. Programming model Couchbase provides client libraries for different programming languages such as Java / .NET / PHP / Ruby / C / Python / Node.js For read, Couchbase provides a key-based lookup mechanism where the client is expected to provide the key, and only the server hosting the data (with that key) will be contacted. Couchbase also provides a query mechanism to retrieve data where the client provides a query (for example, range based on some secondary key) as well as the view (basically the index). The query will be broadcasted to all servers in the cluster and the result will be merged and sent back to the client. For write, Couchbase provides a key-based update mechanism where the client sends in an updated document with the key (as doc id). When handling write request, the server will return to client’s write request as soon as the data is stored in RAM on the active server, which offers the lowest latency for write requests. Following is the core API that Couchbase offers. (in an abstract sense) # Get a document by key doc = get(key) # Modify a document, notice the whole document # need to be passed in set(key, doc) # Modify a document when no one has modified it # since my last read casVersion = doc.getCas() cas(key, casVersion, changedDoc) # Create a new document, with an expiration time # after which the document will be deleted addIfNotExist(key, doc, timeToLive) # Delete a document delete(key) # When the value is an integer, increment the integer increment(key) # When the value is an integer, decrement the integer decrement(key) # When the value is an opaque byte array, append more # data into existing value append(key, newData) # Query the data results = query(viewName, queryParameters) In Couchbase, document is the unit of manipulation. Currently Couchbase doesn't support server-side execution of custom logic. Couchbase server is basically a passive store and unlike other document oriented DB, Couchbase doesn't support field-level modification. In case of modifying documents, client need to retrieve documents by its key, do the modification locally and then send back the whole (modified) document back to the server. This design tradeoff network bandwidth (since more data will be transferred across the network) for CPU (now CPU load shift to client). Couchbase currently doesn't support bulk modification based on a condition matching. Modification happens only in a per document basis. (client will save the modified document one at a time). Transaction Model Similar to many NOSQL databases, Couchbase’s transaction model is primitive as compared to RDBMS. Atomicity is guaranteed at a single document and transactions that span update of multiple documents are unsupported. To provide necessary isolation for concurrent access, Couchbase provides a CAS (compare and swap) mechanism which works as follows … When the client retrieves a document, a CAS ID (equivalent to a revision number) is attached to it. While the client is manipulating the retrieved document locally, another client may modify this document. When this happens, the CAS ID of the document at the server will be incremented. Now, when the original client submits its modification to the server, it can attach the original CAS ID in its request. The server will verify this ID with the actual ID in the server. If they differ, the document has been updated in between and the server will not apply the update. The original client will re-read the document (which now has a newer ID) and re-submit its modification. Couchbase also provides a locking mechanism for clients to coordinate their access to documents. Clients can request a LOCK on the document it intends to modify, update the documents and then releases the LOCK. To prevent a deadlock situation, each LOCK grant has a timeout so it will automatically be released after a period of time. Deployment Architecture In a typical setting, a Couchbase DB resides in a server clusters involving multiple machines. Client library will connect to the appropriate servers to access the data. Each machine contains a number of daemon processes which provides data access as well as management functions. The data server, written in C/C++, is responsible to handle get/set/delete request from client. The Management server, written in Erlang, is responsible to handle the query traffic from client, as well as manage the configuration and communicate with other member nodes in the cluster. Virtual Buckets The basic unit of data storage in Couchbase DB is a JSON document (or primitive data type such as int and byte array) which is associated with a key. The overall key space is partitioned into 1024 logical storage unit called "virtual buckets" (or vBucket). vBucket are distributed across machines within the cluster via a map that is shared among servers in the cluster as well as the client library. High availability is achieved through data replication at the vBucket level. Currently Couchbase supports one active vBucket zero or more standby replicas hosted in other machines. Curremtly the standby server are idle and not serving any client request. In future version of Couchbase, the standby replica will be able to serve read request. Load balancing in Couchbase is achieved as follows: Keys are uniformly distributed based on the hash function When machines are added and removed in the cluster. The administrator can request a redistribution of vBucket so that data are evenly spread across physical machines. Management Server Management server performs the management function and co-ordinate the other nodes within the cluster. It includes the following monitoring and administration functions Heartbeat: A watchdog process periodically communicates with all member nodes within the same cluster to provide Couchbase Server health updates. Process monitor: This subsystem monitors execution of the local data manager, restarting failed processes as required and provide status information to the heartbeat module. Configuration manager: Each Couchbase Server node shares a cluster-wide configuration which contains the member nodes within the cluster, a vBucket map. The configuration manager pull this config from other member nodes at bootup time. Within a cluster, one node’s Management Server will be elected as the leader which performs the following cluster-wide management function Controls the distribution of vBuckets among other nodes and initiate vBucket migration Orchestrates the failover and update the configuration manager of member nodes If the leader node crashes, a new leader will be elected from surviving members in the cluster. When a machine in the cluster has crashed, the leader will detect that and notify member machines in the cluster that all vBuckets hosted in the crashed machine is dead. After getting this signal, machines hosting the corresponding vBucket replica will set the vBucket status as “active”. The vBucket/server map is updated and eventually propagated to the client lib. Notice that at this moment, the replication level of the vBucket will be reduced. Couchbase doesn’t automatically re-create new replicas which will cause data copying traffic. Administrator can issue a command to explicitly initiate a data rebalancing. The crashed machine, after reboot can rejoin the cluster. At this moment, all the data it stores previously will be completely discard and the machine will be treated as a brand new empty machine. As more machines are put into the cluster (for scaling out), vBucket should be redistributed to achieve a load balance. This is currently triggered by an explicit command from the administrator. Once receive the “rebalance” command, the leader will compute the new provisional map which has the balanced distribution of vBuckets and send this provisional map to all members of the cluster. To compute the vBucket map and migration plan, the leader attempts the following objectives: Evenly distribute the number of active vBuckets and replica vBuckets among member nodes. Place the active copy and each replicas in physically separated nodes. Spread the replica vBucket as wide as possible among other member nodes. Minimize the amount of data migration Orchestrate the steps of replica redistribution so no node or network will be overwhelmed by the replica migration. Once the vBucket maps is determined, the leader will pass the redistribution map to each member in the cluster and coordinate the steps of vBucket migration. The actual data transfer happens directly between the origination node to the destination node. Notice that since we have generally more vBuckets than machines. The workload of migration will be evenly distributed automatically. For example, when new machines are added into the clusters, all existing machines will migrate some portion of its vBucket to the new machines. There is no single bottleneck in the cluster. Throughput the migration and redistribution of vBucket among servers, the life cycle of a vBucket in a server will be in one of the following states “Active”: means the server is hosting the vBucket is ready to handle both read and write request “Replica”: means the server is hosting the a copy of the vBucket that may be slightly out of date but can take read request that can tolerate some degree of outdate. “Pending”: means the server is hosting a copy that is in a critical transitional state. The server cannot take either read or write request at this moment. “Dead”: means the server is no longer responsible for the vBucket and will not take either read or write request anymore. Data Server Data server implements the memcached APIs such as get, set, delete, append, prepend, etc. It contains the following key datastructure: One in-memory hashtable (key by doc id) for the corresponding vBucket hosted. The hashtable acts as both a metadata for all documents as well as a cache for the document content. Maintain the entry gives a quick way to detect whether the document exists on disk. To support async write, there is a checkpoint linkedlist per vBucket holding the doc id of modified documents that hasn't been flushed to disk or replicated to the replica. To handle a "GET" request Data server routes the request to the corresponding ep-engine responsible for the vBucket. The ep-engine will lookup the document id from the in-memory hastable. If the document content is found in cache (stored in the value of the hashtable), it will be returned. Otherwise, a background disk fetch task will be created and queued into the RO dispatcher queue. The RO dispatcher then reads the value from the underlying storage engine and populates the corresponding entry in the vbucket hash table. Finally, the notification thread notifies the disk fetch completion to the memcached pending connection, so that the memcached worker thread can revisit the engine to process a get request. To handle a "SET" request, a success response will be returned to the calling client once the updated document has been put into the in-memory hashtable with a write request put into the checkpoint buffer. Later on the Flusher thread will pickup the outstanding write request from each checkpoint buffer, lookup the corresponding document content from the hashtable and write it out to the storage engine. Of course, data can be lost if the server crashes before the data has been replicated to another server and/or persisted. If the client requires a high data availability across different crashes, it can issue a subsequent observe() call which blocks on the condition that the server persist data on disk, or the server has replicated the data to another server (and get its ACK). Overall speaking, the client has various options to tradeoff data integrity with throughput. Hashtable Management To synchronize accesses to a vbucket hash table, each incoming thread needs to acquire a lock before accessing a key region of the hash table. There are multiple locks per vbucket hash table, each of which is responsible for controlling exclusive accesses to a certain ket region on that hash table. The number of regions of a hash table can grow dynamically as more documents are inserted into the hash table. To control the memory size of the hashtable, Item pager thread will monitor the memory utilization of the hashtable. Once a high watermark is reached, it will initiate an eviction process to remove certain document content from the hashtable. Only entries that is not referenced by entries in the checkpoint buffer can be evicted because otherwise the outstanding update (which only exists in hashtable but not persisted) will be lost. After eviction, the entry of the document still remains in the hashtable; only the document content of the document will be removed from memory but the metadata is still there. The eviction process stops after reaching the low watermark. The high / low water mark is determined by the bucket memory quota. By default, the high water mark is set to 75% of bucket quota, while the low water mark is set to 60% of bucket quota. These water marks can be configurable at runtime. In CouchDb, every document is associated with an expiration time and will be deleted once it is expired. Expiry pager is responsible for tracking and removing expired document from both the hashtable as well as the storage engine (by scheduling a delete operation). Checkpoint Manager Checkpoint manager is responsible to recycle the checkpoint buffer, which holds the outstanding update request, consumed by the two downstream processes, Flusher and TAP replicator. When all the request in the checkpoint buffer has been processed, the checkpoint buffer will be deleted and a new one will be created. TAP Replicator TAP replicator is responsible to handle vBucket migration as well as vBucket replication from active server to replica server. It does this by propagating the latest modified document to the corresponding replica server. At the time a replica vBucket is established, the entire vBucket need to be copied from the active server to the empty destination replica server as follows The in-memory hashtable at the active server will be transferred to the replica server. Notice that during this period, some data may be updated and therefore the data set transfered to the replica can be inconsistent (some are the latest and some are outdated). Nevertheless, all updates happen after the start of transfer is tracked in the checkpoint buffer. Therefore, after the in-memory hashtable transferred is completed, the TAP replicator can pickup those updates from the checkpoint buffer. This ensures the latest versioned of changed documents are sent to the replica, and hence fix the inconsistency. However the hashtable cache doesn’t contain all the document content. Data also need to be read from the vBucket file and send to the replica. Notice that during this period, update of vBucket will happen in active server. However, since the file is appended only, subsequent data update won’t interfere the vBucket copying process. After the replica server has caught up, subsequent update at the active server will be available at its checkpoint buffer which will be pickup by the TAP replicator and send to the replica server. CouchDB Storage Structure Data server defines an interface where different storage structure can be plugged-in. Currently it supports both a SQLite DB as well as CouchDB. Here we describe the details of CouchDb, which provides a super high performance storage mechanism underneath the Couchbase technology. Under the CouchDB structure, there will be one file per vBucket. Data are written to this file in an append-only manner, which enables Couchbase to do mostly sequential writes for update, and provide the most optimized access patterns for disk I/O. This unique storage structure attributes to Couchbase’s fast on-disk performance for write-intensive applications. The following diagram illustrate the storage model and how it is modified by 3 batch updates (notice that since updates are asynchronous, it is perform by "Flusher" thread in batches). The Flusher thread works as follows: 1) Pick up all pending write request from the dirty queue and de-duplicate multiple update request to the same document. 2) Sort each request (by key) into corresponding vBucket and open the corresponding file 3) Append the following into the vBucket file (in the following contiguous sequence) All document contents in such write request batch. Each document will be written as [length, crc, content] one after one sequentially. The index that stores the mapping from document id to the document’s position on disk (called the BTree by-id) The index that stores the mapping from update sequence number to the document’s position on disk. (called the BTree by-seq) The by-id index plays an important role for looking up the document by its id. It is organized as a B-Tree where each node contains a key range. To lookup a document by id, we just need to start from the header (which is the end of the file), transfer to the root BTree node of the by-id index, and then further traverse to the leaf BTree node that contains the pointer to the actual document position on disk. During the write, the similar mechanism is used to trace back to the corresponding BTree node that contains the id of the modified documents. Notice that in the append-only model, update is not happening in-place, instead we located the existing location and copy it over by appending. In other words, the modified BTree node will be need to be copied over and modified and finally paste to the end of file, and then its parent need to be modified to point to the new location, which triggers the parents to be copied over and paste to the end of file. Same happens to its parents’ parent and eventually all the way to the root node of the BTree. The disk seek can be at the O(logN) complexity. The by-seq index is used to keep track of the update sequence of lived documents and is used for asynchronous catchup purposes. When a document is created, modified or deleted, a sequence number is added to the by-seq btree and the previous seq node will be deleted. Therefore, for cross-site replication, view index update and compaction, we can quickly locate all the lived documents in the order of their update sequence. When a vBucket replicator asks for the list of update since a particular time, it provides the last sequence number in previous update, the system will then scan through the by-seq BTree node to locate all the document that has sequence number larger than that, which effectively includes all the document that has been modified since the last replication. As time goes by, certain data becomes garbage (see the grey-out region above) and become unreachable in the file. Therefore, we need a garbage collection mechanism to clean up the garbage. To trigger this process, the by-id and by-seq B-Tree node will keep track of the data size of lived documents (those that is not garbage) under its substree. Therefore, by examining the root BTree node, we can determine the size of all lived documents within the vBucket. When the ratio of actual size and vBucket file size fall below a certain threshold, a compaction process will be triggered whose job is to open the vBucket file and copy the survived data to another file. Technically, the compaction process opens the file and read the by-seq BTree at the end of the file. It traces the Btree all the way to the leaf node and copy the corresponding document content to the new file. The compaction process happens while the vBucket is being updated. However, since the file is appended only, new changes are recorded after the BTree root that the compaction has opened, so subsequent data update won’t interfere with the compaction process. When the compaction is completed, the system need to copy over the data that was appended since the beginning of the compaction to the new file. View Index Structure Unlike most indexing structure which provide a pointer from the search attribute back to the document. The CouchDb index (called View Index) is better perceived as a denormalized table with arbitrary keys and values loosely associated to the document. Such denormalized table is defined by a user-provided map() and reduce() function. map = function(doc) { … emit(k1, v1) … emit(k2, v2) … } reduce = function(keys, values, isRereduce) { if (isRereduce) { // Do the re-reduce only on values (keys will be null) } else { // Do the reduce on keys and values } // result must be ready for input values to re-reduce return result } Whenever a document is created, updated, deleted, the corresponding map(doc) function will be invoked (in an asynchronous manner) to generate a set of key/value pairs. Such key/value will be stored in a B-Tree structure. All the key/values pairs of each B-Tree node will be passed into the reduce() function, which compute an aggregated value within that B-Tree node. Re-reduce also happens in non-leaf B-Tree nodes which further aggregate the aggregated value of child B-Tree nodes. The management server maintains the view index and persisted it to a separate file. Create a view index is perform by broadcast the index creation request to all machines in the cluster. The management process of each machine will read its active vBucket file and feed each surviving document to the Map function. The key/value pairs emitted by the Map function will be stored in a separated BTree index file. When writing out the BTree node, the reduce() function will be called with the list of all values in the tree node. Its return result represent a partially reduced value is attached to the BTree node. The view index will be updated incrementally as documents are subsequently getting into the system. Periodically, the management process will open the vBucket file and scan all documents since the last sequence number. For each changed document since the last sync, it invokes the corresponding map function to determine the corresponding key/value into the BTree node. The BTree node will be split if appropriate. Underlying, Couchbase use a back index to keep track of the document with the keys that it previously emitted. Later when the document is deleted, it can look up the back index to determine what those key are and remove them. In case the document is updated, the back index can also be examined; semantically a modification is equivalent to a delete followed by an insert. The following diagram illustrates how the view index file will be incrementally updated via the append-only mechanism. Query Processing Query in Couchbase is made against the view index. A query is composed of the view name, a start key and end key. If the reduce() function isn’t defined, the query result will be the list of values sorted by the keys within the key range. In case the reduce() function is defined, the query result will be a single aggregated value of all keys within the key range. If the view has no reduce() function defined, the query processing proceeds as follows: Client issue a query (with view, start/end key) to the management process of any server (unlike a key based lookup, there is no need to locate a specific server). The management process will broadcast the request to other management process on all servers (include itself) within the cluster. Each management process (after receiving the broadcast request) do a local search for value within the key range by traversing the BTree node of its view file, and start sending back the result (automatically sorted by the key) to the initial server. The initial server will merge the sorted result and stream them back to the client. However, if the view has reduce() function defined, the query processing will involve computing a single aggregated value as follows: Client issue a query (with view, start/end key) to the management process of any server (unlike a key based lookup, there is no need to locate a specific server). The management process will broadcast the request to other management process on all servers (include itself) within the cluster. Each management process do a local reduce for value within the key range by traversing the BTree node of its view file to compute the reduce value of the key range. If the key range span across a BTree node, the pre-computed of the sub-range can be used. This way, the reduce function can reuse a lot of partially reduced values and doesn’t need to recomputed every value of the key range from scratch. The original server will do a final re-reduce() in all the return value from each other servers, and then passed back the final reduced value to the client. To illustrate the re-reduce concept, lets say the query has its key range from A to F. Instead of calling reduce([A,B,C,D,E,F]), the system recognize the BTree node that contains [B,C,D] has been pre-reduced and the result P is stored in the BTree node, so it only need to call reduce(A,P,E,F). Update View Index as vBucket migrates Since the view index is synchronized with the vBuckets in the same server, when the vBucket has migrated to a different server, the view index is no longer correct; those key/value that belong to a migrated vBucket should be discarded and the reduce value cannot be used anymore. To keep track of the vBucket and key in the view index, each bTree node has a 1024-bitmask indicating all the vBuckets that is covered in the subtree (ie: it contains a key emitted from a document belonging to the vBucket). Such bit-mask is maintained whenever the bTree node is updated. At the server-level, a global bitmask is used to indicate all the vBuckets that this server is responsible for. In processing the query of the map-only view, before the key/value pair is returned, an extra check will be perform for each key/value pair to make sure its associated vBucket is what this server is responsible for. When processing the query of a view that has a reduce() function, we cannot use the pre-computed reduce value if the bTree node contains a vBucket that the server is not responsible for. In this case, the bTree node’s bit mask is compared with the global bit mask. In case if they are not aligned, then the reduce value need to be recomputed. Here is an example to illustrate this process Couchbase is one of the popular NOSQL technology built on a solid technology foundation designed for high performance. In this post, we have examined a number of such key features: Load balancing between servers inside a cluster that can grow and shrink according to workload conditions. Data migration can be used to re-achieve workload balance. Asynchronous write provides lowest possible latency to client as it returns once the data is store in memory. Append-only update model pushes most update transaction into sequential disk access, hence provide extremely high throughput for write intensive applications. Automatic compaction ensures the data lay out on disk are kept optimized all the time. Map function can be used to pre-compute view index to enable query access. Summary data can be pre-aggregated using the reduce function. Overall, this cut down the workload of query processing dramatically. For a review on NOSQL architecture in general and some theoretical foundation, I have wrote a NOSQL design pattern blog, as well as some fundamental difference between SQL and NOSQL. For other NOSQL technologies, please read my other blog on MongoDb, Cassandra and HBase, Memcached Special thanks to Damien Katz and Frank Weigel from Couchbase team who provide a lot of implementation details of Couchbase.

In my last post, I discussed some of the key considerations when moving an application to the cloud. To provide a better understanding, I’m using a simple scenario-based example to illustrate how an application could be moved to the cloud. This article will explain the challenges a company might face, the current architecture of the example application, and finally what the company should expect when moving an application to the cloud. My next article will discuss the recommended solution in more detail. Disclaimer Company name, logo, business, scenario, and incidents either are used fictitiously. Any resemblance to an actual company is entirely coincidental. Background Idelma is a ticket selling provider that sells tickets to concerts, sports event, and music gigs. Tickets are sold offline through ticket counters and online through a website called TicketOnline. Customers visiting TicketOnline can browse list of available shows, find out more information on each show, and finally purchase tickets online. When a ticket is purchased, it’s reserved but will not be processed immediately. Other processes such as generating ticket and sending the generated ticket along with the receipt will be done asynchronously in a few minutes time. Current Challenges During peak season (typically in July and December), TicketOnline suffered from heavy traffic that caused slow response time. The traffic for off-peak season is normally about 100,000 to 200,000 hits per day, with the average of 8 to 15 on-going shows. In peak season, the traffic may reach five to seven times more than off-peak season. The following diagram illustrates the web server hits counter of TicketOnline over the last three years. Figure 1 – TicketOnline web server hits counter for the last three years Additionally, the current infrastructure setup is not designed to be highly-available. This results in several periods of downtime each year. The options: on-premise vs cloud Idelma’s IT Manager Mr. Anthony recognizes the issues and decides to make some improvement to bring better competitive advantages to the company. When reading an article online, he discovered that cloud computing may be a good solution to address the issues. Another option would be to purchase a more powerful set of hardware that could handle the load. With that, he has done a pros and cons analysis of the two options: On-premise hardware investment There are at least two advantages of investing in more hardware. One, they will have full control over the infrastructure, and can use the server for other purposes when necessary. Second, there might be less or no modification needed on the application at all, depending on how it is architected and designed. If they decide to scale up (vertically), they might not need to make any changes. However, if they decide to scale out (horizontally) to a web farm model, a re-design would be needed. On the other hand, there are also several disadvantages of on-premise hardware investment. For sure, upfront investment in purchasing hardware and software are considered relatively expensive. Next, they would need to be able to answer the following questions: How much hardware and software should be purchased? What are the hardware specifications? If the capacity planning is not properly done, it may lead to either a waste of capacity or insufficient of capacity. Another concern is, when adding more hardware, more manpower might be needed as well. Cloud For cloud computing, there’s almost no upfront investment required for hardware, and in some cases software doesn’t pose a large upfront cost either. Another advantage is the cloud’s elastic nature fits TicketOnline periodic bursting very much. Remember, they face high load only in June and December. Another advantage would be less responsibility. The administrator can have more time to focus on managing the application since the infrastructure is managed by the provider. Though there are a number of advantages, there are also some disadvantages when choosing a cloud platform. For one thing, they might have less control over the infrastructure. As discussed in the previous article, there might also be some architectural changes when moving an application to the cloud. However, these can be dealt with in a one-time effort. The figure below summarizes the considerations between the two options: Figure 2 – Considerations of an On-premise or Cloud solution After looking at his analysis, Mr. Anthony believes that the cloud will bring more competitive advantages to the company. Understanding that Windows Azure offers various services for building internet-scale application, and Idelma is also an existing Microsoft customer, Mr. Anthony decided to explore Windows Azure. After evaluating the pricing, he is even more comfortable to step ahead. Quick preview of the current system Now, let’s take a look of the current architecture of TicketOnline. Figure 3 – TicketOnline Current Architecture TicketOnline web application The web application is hosted on a single instance physical server. It is running on Windows Server 2003 R2 as operating system with Internet Information Services (IIS) 6 as the web server and ASP.NET 2.0 as the web application framework. Database SQL Server 2005 is used as database engine to store mainly relational data for the application. Additionally, it is also used to store logs such as trace logs, performance-counters logs, and IIS logs. File server Unstructured files such as images and documents are stored separately in a file server. Interfacing with another system The application would need to interface with a proprietary CRM system that runs on a dedicated server to retrieve customer profiles through asmx web service. Batch Job As mentioned previously, receipt and ticket generation will happen asynchronously after purchasing is made. A scheduler-based batch job will perform asynchronous tasks every 10 minutes. The tasks include verifying booking details, generating tickets, and sending the ticket along with the receipt as an email to customer. The intention of an asynchronous process is to minimize concurrent access load as much as possible. This batch job is implemented as a Windows Service installed in a separated server. SMTP Server On-premise SMTP Server will be used to send email, initiated either from the batch job engine or the web application. Requirements for migration The application should be migrated to the cloud with the following requirements: The customer expects a cost effective solution in terms of the migration effort as well as the monthly running cost. There aren’t any functional changes on the system. Meaning, the user (especially front-end user) should not see any differences in term of functionality. As per policy, this propriety CRM system will not be moved to the cloud. The web service consumption should be consumed in secured manner. Calling for partners As the in-house IT team does not have competency and experience with Windows Azure, Mr. Anthony contacted Microsoft to suggest a partner who is capable to deliver the migration. Before a formal request for proposal (RFP) is made, he expects partner to provide the following: High-level architecture diagram how the system will look when moving to the cloud. Explanation of each component illustrated on the diagram. The migration processes, effort required, and potential challenges. If Microsoft recommends you as the partner, how will you handle this case? What will the architecture look like in your proposed solution? The most exciting part will come in the next article when I go into more detail on which solution is recommended and how the migration process takes place.

Performance problems are one of the biggest challenges to expect when designing and implementing Java EE related technologies.

infographic: clouds computing and history i have prepared three new infographics for you;aall of them related with cloud computing. these infographics will tell you about history of cloud computing, its definition, and who needs this cloud. i think that this will be interesting for you. information graphics (known as infographics) are one of the best ways to transfer some information into a reader’s mind. it can be something new, or other useful information gathered in one place. nowadays many people don’t have enough time to read a lot of text on multiple screens. infographics makes the information intuitive and understandable. that’s why we would like to share the best relevant infographics from all over the web. original source: cloud computing by the small business authority original source: a complete history of cloud computing original source: hosting decisions, from the chalkboard

Some time ago I came across What's the purpose of AsyncContext.start(...) in Servlet 3.0? question. Quoting the Javadoc of aforementioned method: Causes the container to dispatch a thread, possibly from a managed thread pool, to run the specified Runnable. To remind all of you, AsyncContext is a standard way defined in Servlet 3.0 specification to handle HTTP requests asynchronously. Basically HTTP request is no longer tied to an HTTP thread, allowing us to handle it later, possibly using fewer threads. It turned out that the specification provides an API to handle asynchronous threads in a different thread pool out of the box. First we will see how this feature is completely broken and useless in Tomcat and Jetty - and then we will discuss why the usefulness of it is questionable in general. Our test servlet will simply sleep for given amount of time. This is a scalability killer in normal circumstances because even though sleeping servlet is not consuming CPU, but sleeping HTTP thread tied to that particular request consumes memory - and no other incoming request can use that thread. In our test setup I limited the number of HTTP worker threads to 10 which means only 10 concurrent requests are completely blocking the application (it is unresponsive from the outside) even though the application itself is almost completely idle. So clearly sleeping is an enemy of scalability. @WebServlet(urlPatterns = Array("/*")) class SlowServlet extends HttpServlet with Logging { protected override def doGet(req: HttpServletRequest, resp: HttpServletResponse) { logger.info("Request received") val sleepParam = Option(req.getParameter("sleep")) map {_.toLong} TimeUnit.MILLISECONDS.sleep(sleepParam getOrElse 10) logger.info("Request done") } } Benchmarking this code reveals that the average response times are close to sleep parameter as long as the number of concurrent connections is below the number of HTTP threads. Unsurprisingly the response times begin to grow the moment we exceed the HTTP threads count. Eleventh connection has to wait for any other request to finish and release worker thread. When the concurrency level exceeds 100, Tomcat begins to drop connections - too many clients are already queued. So what about the the fancy AsyncContext.start() method (do not confuse with ServletRequest.startAsync())? According to the JavaDoc I can submit any Runnable and the container will use some managed thread pool to handle it. This will help partially as I no longer block HTTP worker threads (but still another thread somewhere in the servlet container is used). Quickly switching to asynchronous servlet: @WebServlet(urlPatterns = Array("/*"), asyncSupported = true) class SlowServlet extends HttpServlet with Logging { protected override def doGet(req: HttpServletRequest, resp: HttpServletResponse) { logger.info("Request received") val asyncContext = req.startAsync() asyncContext.setTimeout(TimeUnit.MINUTES.toMillis(10)) asyncContext.start(new Runnable() { def run() { logger.info("Handling request") val sleepParam = Option(req.getParameter("sleep")) map {_.toLong} TimeUnit.MILLISECONDS.sleep(sleepParam getOrElse 10) logger.info("Request done") asyncContext.complete() } }) } } We are first enabling the asynchronous processing and then simply moving sleep() into a Runnable and hopefully a different thread pool, releasing the HTTP thread pool. Quick stress test reveals slightly unexpected results (here: response times vs. number of concurrent connections): Guess what, the response times are exactly the same as with no asynchronous support at all (!) After closer examination I discovered that when AsyncContext.start() is called Tomcat submits given task back to... HTTP worker thread pool, the same one that is used for all HTTP requests! This basically means that we have released one HTTP thread just to utilize another one milliseconds later (maybe even the same one). There is absolutely no benefit of calling AsyncContext.start() in Tomcat. I have no idea whether this is a bug or a feature. On one hand this is clearly not what the API designers intended. The servlet container was suppose to manage separate, independent thread pool so that HTTP worker thread pool is still usable. I mean, the whole point of asynchronous processing is to escape the HTTP pool. Tomcat pretends to delegate our work to another thread, while it still uses the original worker thread pool. So why I consider this to be a feature? Because Jetty is "broken" in exactly same way... No matter whether this works as designed or is only a poor API implementation, using AsyncContext.start() in Tomcat and Jetty is pointless and only unnecessarily complicates the code. It won't give you anything, the application works exactly the same under high load as if there was no asynchronous logic at all. But what about using this API feature on correct implementations like IBM WAS? It is better, but still the API as is doesn't give us much in terms of scalability. To explain again: the whole point of asynchronous processing is the ability to decouple HTTP request from an underlying thread, preferably by handling several connections using the same thread. AsyncContext.start() will run the provided Runnable in a separate thread pool. Your application is still responsive and can handle ordinary requests while long-running request that you decided to handle asynchronously are processed in a separate thread pool. It is better, unfortunately the thread pool and thread per connection idiom is still a bottle-neck. For the JVM it doesn't matter what type of threads are started - they still occupy memory. So we are no longer blocking HTTP worker threads, but our application is not more scalable in terms of concurrent long-running tasks we can support. In this simple and unrealistic example with sleeping servlet we can actually support thousand of concurrent (waiting) connections using Servlet 3.0 asynchronous support with only one extra thread - and without AsyncContext.start(). Do you know how? Hint: ScheduledExecutorService. Postscriptum: Scala goodness I almost forgot. Even though examples were written in Scala, I haven't used any cool language features yet. Here is one: implicit conversions. Make this available in your scope: implicit def blockToRunnable[T](block: => T) = new Runnable { def run() { block } } And suddenly you can use code block instead of instantiating Runnable manually and explicitly: asyncContext start { logger.info("Handling request") val sleepParam = Option(req.getParameter("sleep")) map { _.toLong} TimeUnit.MILLISECONDS.sleep(sleepParam getOrElse 10) logger.info("Request done") asyncContext.complete() } Sweet!

First blogged about this on my previous blog site here: http://consultingblogs.emc.com/merrickchaffer/archive/2011/02/14/virtualization-in-wpf-with-virtualizingstackpanel.aspx However, having come across this again today on a project, I thought it was important enough to re-blog! Finally managed to figure out how to get virtualization to actually behave itself in a listbox wpf control. Turns out that in order for Virtualization to work, you need three things satisfied. Use a control that supports virtualization (e.g. list box or list view). (see Controls That Implement Performance Features section at bottom of this page for more info http://msdn.microsoft.com/en-us/library/cc716879.aspx#Controls ) Ensure that the ScrollViewer.CanContentScroll attached property is set to True on the containing list box / list view control. Ensure that either the list box has a height set, or that it is contained within a parent Grid row, where that row definition has a height set (Height="*" will do if you want it to occupy the Client window height). Note: Do not use height=”Auto” as this will not work, as this instructs WPF to simply size the row to the height needed to fit all the items of the list box in, hence you do not get the vertical scroll bar appearing. Ensure that there is no wrapping ScrollViewer control around the list box, as this will prevent virtualization from occuring. Ensure that you use a VirtualizingStackPanel in the ItemsPanelTemplate for the ListBox.ItemsPanel Example

A few weeks ago, I co-authored an article (with my colleague Rama Ramani) about how the Screen Actors Guild Awards website migrated its Drupal deployment from LAMP to Windows Azure: Azure Real World: Migrating a Drupal Site from LAMP to Windows Azure. Since then, Rama and another colleague, Jason Roth, have been working on writing up how the SAG Awards website was managed and monitored in Windows Azure. The article below is the fruit of their work…a very interesting/educational read. Overview Drupal is an open source content management system that runs on PHP. Windows Azure offers a flexible platform for hosting, managing, and scaling Drupal deployments. This paper focuses on an approach to host Drupal sites on Windows Azure, based on learning from a BPD Customer Programs Design Win engagement with the Screen Actors Guild Awards Drupal website. This paper covers guidelines and best practices for managing an existing Drupal web site in Windows Azure. For more information on how to migrate Drupal applications to Windows Azure, see Azure Real World: Migrating a Drupal Site from LAMP to Windows Azure. The target audience for this paper is Drupal administrators who have some exposure to Windows Azure. More detailed pointers to Windows Azure content is provided throughout the paper as links. Drupal Application Architecture on Windows Azure Before reviewing the management and monitoring guidelines, it is important to understand the architecture of a typical Drupal deployment on Windows Azure. First, the following diagram displays the basic architecture of Drupal running on Windows and IIS7. In the Windows Server scenario, you could have one or more machines hosting the web site in a farm. Those machines would either persist the site content to the file system or point to other network shares. For Windows Azure, the basic architecture is the same, but there are some differences. In Windows Azure the site is hosted on a web role. A web role instance is hosted on a Windows Server 2008 virtual machine within the Windows Azure datacenter. Like the web farm, you can have multiple instances running the site. But there is no persistence guarantee for the data on the file system. Because of this, much of the shared site content should be stored in Windows Azure Blob storage. This allows them to be highly available and durable. Usually, a large portion of the site caters to static content which lends well to caching. And caching can be applied in a set of places – browser level caching, CDN to cache content in the edge closer to the browser clients, caching in Azure to reduce the load on backend, etc. Finally, the database can be located in SQL Azure. The following diagram shows these differences. For monitoring and management, we will look at Drupal on Windows Azure from three perspectives: Availability: Ensure the web site does not go down and that all tiers are setup correctly. Apply best practices to ensure that the site is deployed across data centers and perform backup operations regularly. Scalability: Correctly handle changes in user load. Understand the performance characteristics of the site. Manageability: Correctly handle updates. Make code and site changes with no downtime when possible. Although some management tasks span one or more of these categories, it is still helpful to discuss Drupal management on Windows Azure within these focus areas. Availability One main goal is that the Drupal site remains running and accessible to all end-users. This involves monitoring both the site and the SQL Azure database that the site depends on. In this section, we will briefly look at monitoring and backup tasks. Other crossover areas that affect availability will be discussed in the next section on scalability. Monitoring With any application, monitoring plays an important role with managing availability. Monitoring data can reveal whether users are successfully using the site or whether computing resources are meeting the demand. Other data reveals error counts and possibly points to issues in a specific tier of the deployment. There are several monitoring tools that can be used. The Windows Azure Management Portal. Windows Azure diagnostic data. Custom monitoring scripts. System Center Operations Manager. Third party tools such as Azure Diagnostics Manager and Azure Storage Explorer. The Windows Azure Management Portal can be used to ensure that your deployments are successful and running. You can also use the portal to manage features such as Remote Desktop so that you can directly connect to machines that are running the Drupal site. Windows Azure diagnostics allows you to collect performance counters and logs off of the web role instances that are running the Drupal site. Although there are many options for configuring diagnostics in Azure, the best solution with Drupal is to use a diagnostics configuration file. The following configuration file demonstrates some basic performance counters that can monitor resources such as memory, processor utilization, and network bandwidth. For more information about setting up diagnostic configuration files, see How to Use the Windows Azure Diagnostics Configuration File. This information is stored locally on each role instance and then transferred to Windows Azure storage per a defined schedule or on-demand. See Getting Started with Storing and Viewing Diagnostic Data in Windows Azure Storage. Various monitoring tools, such as Azure Diagnostics Manager, help you to more easily analyze diagnostic data. Monitoring the performance of the machines hosting the Drupal site is only part of the story. In order to plan properly for both availability and scalability, you should also monitor site traffic, including user load patterns and trends. Standard and custom diagnostic data could contribute to this, but there are also third-party tools that monitor web traffic. For example, if you know that spikes occur in your application during certain days of the week, you could make changes to the application to handle the additional load and increase the availability of the Drupal solution. Backup Tasks To remain highly available, it is important to backup your data as a defense-in-depth strategy for disaster recovery. This is true even though SQL Azure and Windows Azure Storage both implement redundancy to prevent data loss. One obvious reason is that these services cannot prevent administrator error if data is accidentally deleted or incorrectly changed. SQL Azure does not currently have a formal backup technology, although there are many third-party tools and solutions that provide this capability. Usually the database size for a Drupal site is relatively small. In the case of SAG Awards, it was only ~100-150 MB. So performing an entire backup using any strategy was relatively fast. If your database is much larger, you might have to test various backup strategies to find the one that works best. Apart from third-party SQL Azure backup solutions, there are several strategies for obtaining a backup of your data: · Use the Drush tool and the portabledb-export command. · Periodically copy the database using the CREATE DATABASE Transact-SQL command. · Use Data-tier applications (DAC) to assist with backup and restore of the database. SQL Azure backup and data security techniques are described in more detail in the topic, Business Continuity in SQL Azure. Note that bandwidth costs accrue with any backup operation that transfers information outside of the Windows Azure datacenter. To reduce costs, you can copy the database to a database within the same datacenter. Or you can export the data-tier applications to blob storage in the same datacenter. Another potential backup task involves the files in Blob storage. If you keep a master copy of all media files uploaded to Blob storage, then you already have an on-premises backup of those files. However, if multiple administrators are loading files into Blob storage for use on the Drupal site, it is a good idea to enumerate the storage account and to download any new files to a central location. The following PHP script demonstrates how this can be done by backing up all files in Blob storage after a specified modification date. setProxy(true, 'YOUR_PROXY_IF_NEEDED', 80); $blobs = (array)$blobObj->listBlobs(AZURE_STORAGE_CONTAINER, '', '', 35000); backupBlobs($blobs, $blobObj); function backupBlobs($blobs, $blobObj) { foreach ($blobs as $blob) { if (strtotime($blob->lastmodified) >= DEFAULT_BACKUP_FROM_DATE && strtotime($blob->lastmodified) <= DEFAULT_BACKUP_TO_DATE) { $path = pathinfo($blob->name); if ($path['basename'] != '$$$.$$$') { $dir = $path['dirname']; $oldDir = getcwd(); if (handleDirectory($dir)) { chdir($dir); $blobObj->getBlob( AZURE_STORAGE_CONTAINER, $blob->name, $path['basename'] ); chdir($oldDir); } } } } } function handleDirectory($dir) { if (!checkDirExists($dir)) { return mkdir($dir, 0755, true); } return true; } function checkDirExists($dir) { if(file_exists($dir) && is_dir($dir)) { return true; } return false; } ?> This script has a dependency on the Windows Azure SDK for PHP. Also note there are several parameters that you must modify such as the storage account, secret, and backup location. As with SQL Azure, bandwidth and transaction charges apply to a backup script like this. Scalability Drupal sites on Windows Azure can scale as load increased through typical strategies of scale-up, scale-out, and caching. The following sections describe the specifics of how these strategies are implemented in Windows Azure. Typically you make scalability decisions based on monitoring and capacity planning. Monitoring can be done in staging during testing or in production with real-time load. Capacity planning factors in projections for changes in user demand. Scale Up When you configure your web role prior to deployment, you have the option of specifying the Virtual Machine (VM) size, such as Small or ExtraLarge. Each size tier adds additional memory, processing power, and network bandwidth to each instance of your web role. For cost efficiency and smaller units of scale, you can test your application under expected load to find the smallest virtual machine size that meets your requirements. The workload usually in most popular Drupal websites can be separated out into a limited set of Drupal admins making content changes and a large user base who perform mostly read-only workload. End users can be allowed to make ‘writes’, such as uploading blogs or posting in forums, but those changes are not ‘content changes’. Drupal admins are setup to operate without caching so that the writes are made directly to SQL Azure or the corresponding backend database. This workload performs well with Large or ExtraLarge VM sizes. Also, note that the VM size is closely tied to all hardware resources, so if there are many content-rich pages that are streaming content, then the VM size requirements are higher. To make changes to the Virtual Machine size setting, you must change the vmsize attribute of the WebRole element in the service definition file, ServiceDefinition.csdef. A virtual machine size change requires existing applications to be redeployed. Scale Out In addition to the size of each web role instance, you can increase or decrease the number of instances that are running the Drupal site. This spreads the web requests across more servers, enabling the site to handle more users. To change the number of running instances of your web role, see How to Scale Applications by Increasing or Decreasing the Number of Role Instances. Note that some configuration changes can cause your existing web role instances to recycle. You can choose to handle this situation by applying the configuration change and continue running. This is done by handling the RoleEnvironment.Changing event. For more information see, How to Use the RoleEnvironment.Changing Event. A common question for any Windows Azure solution is whether there is some type of built-in automatic scaling. Windows Azure does not provide a service that provides auto-scaling. However, it is possible to create a custom solution that scales Azure services using the Service Management API. For an example of this approach, see An Auto-Scaling Module for PHP Applications in Windows Azure. Caching Caching is an important strategy for scaling Drupal applications on Windows Azure. One reason for this is that SQL Azure implements throttling mechanisms to regulate the load on any one database in the cloud. Code that uses SQL Azure should have robust error handling and retry logic to account for this. For more information, see Error Messages (SQL Azure Database). Because of the potential for load-related throttling as well as for general performance improvement, it is strongly recommended to use caching. Although Windows Azure provides a Caching service, this service does not currently have interoperability with PHP. Because of this, the best solution for caching in Drupal is to use a module that uses an open-source caching technology, such as Memcached. Outside of a specific Drupal module, you can also configure Memcached to work in PHP for Windows Azure. For more information, see Running Memcached on Windows Azure for PHP. Here is also an example of how to get Memcached working in Windows Azure using a plugin: Windows Azure Memcached plugin. In a future paper, we hope to cover this architecture in more detail. For now, here are several design and management considerations related to caching. Area Consideration Design and Implementation For a technology like Memcached, will the cache be collocated (spread across all web role instances)? Or will you attempt to setup a dedicated cache ring with worker roles that only run Memcached? Configuration What memory is required and how will items in the cache be invalidated? Performance and Monitoring What mechanisms will be used to detect the performance and overall health of the cache? For ease of use and cost savings, collocation of the cache across the web role instances of the Drupal site works best. However, this assumes that there is available reserve memory on each instance to apply toward caching. It is possible to increase the virtual machine size setting to increase the amount of available memory on each machine. It is also possible to add additional web role instances to add to the overall memory of the cache while at the same time improving the ability of the web site to respond to load. It is possible to create a dedicated cache cluster in the cloud, but the steps for this are beyond the scope of this paper[RR1] . For Windows Azure Blob storage, there is also a caching feature built into the service called the Content Delivery Network (CDN). CDN provides high-bandwidth access to files in Blob storage by caching copies of the files in edge nodes around the world. Even within a single geographic region, you could see performance improvements as there are many more edge nodes than Windows Azure datacenters. For more information, see Delivering High-Bandwidth Content with the Windows Azure CDN. Manageability It is important to note that each hosted service has a Staging environment and a Production environment. This can be used to manage deployments, because you can load and test and application in staging before performing a VIP swap with production. From a manageability standpoint, Drupal has an advantage on Windows Azure in the way that site content is stored. Because the data necessary to serve pages is stored in the database and blob storage, there is no need to redeploy the application to change the content of the site. Another best practice is to use a separate storage account for diagnostic data than the one that is used for the application itself. This can improve performance and also helps to separate the cost of diagnostic monitoring from the cost of the running application. As mentioned previously, there are several tools that can assist with managing Windows Azure applications. The following table summarizes a few of these choices. Tool Description Windows Azure Management Portal The web interface of the Windows Azure management portal shows deployments, instance counts and properties, and supports many different common management and monitoring tasks. Azure Diagnostics Managerq[RR2] [JR3] A Red Gate Software product that provides advanced monitoring and management of diagnostic data. This tool can be very useful for easily analyzing the performance of the Drupal site to determine appropriate scaling decisions. Azure Storage Explorer A tool created by Neudesic for viewing Windows Azure storage account. This can be useful for viewing both diagnostic data and the files in Blob storage.