helix-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ka...@apache.org
Subject [2/3] [HELIX-404] Update website for 0.6.3 release
Date Thu, 27 Mar 2014 23:08:48 GMT
http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/recipes/rsync_replicated_file_store.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/recipes/rsync_replicated_file_store.md b/website/0.6.3/src/site/markdown/recipes/rsync_replicated_file_store.md
new file mode 100644
index 0000000..899387f
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/recipes/rsync_replicated_file_store.md
@@ -0,0 +1,162 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+Near-Realtime Rsync Replicated File System
+------------------------------------------
+
+### Quick Demo
+
+* This demo starts 3 instances with id's as ```localhost_12001, localhost_12002, localhost_12003```
+* Each instance stores its files under ```/tmp/<id>/filestore```
+* ```localhost_12001``` is designated as the master, and ```localhost_12002``` and ```localhost_12003``` are the slaves
+* Files written to the master are replicated to the slaves automatically. In this demo, a.txt and b.txt are written to ```/tmp/localhost_12001/filestore``` and they get replicated to other folders.
+* When the master is stopped, ```localhost_12002``` is promoted to master.
+* The other slave ```localhost_12003``` stops replicating from ```localhost_12001``` and starts replicating from new master ```localhost_12002```
+* Files written to new master ```localhost_12002``` are replicated to ```localhost_12003```
+* In the end state of this quick demo, ```localhost_12002``` is the master and ```localhost_12003``` is the slave. Manually create files under ```/tmp/localhost_12002/filestore``` and see that appear in ```/tmp/localhost_12003/filestore```
+* Ignore the interrupted exceptions on the console :-)
+
+
+```
+git clone https://git-wip-us.apache.org/repos/asf/helix.git
+cd helix
+git checkout tags/helix-0.6.3
+cd recipes/rsync-replicated-file-system/
+mvn clean install package -DskipTests
+cd target/rsync-replicated-file-system-pkg/bin
+chmod +x *
+./quickdemo
+
+```
+
+### Overview
+
+There are many applications that require storage for storing large number of relatively small data files. Examples include media stores to store small videos, images, mail attachments etc. Each of these objects is typically kilobytes, often no larger than a few megabytes. An additional distinguishing feature of these use cases is that files are typically only added or deleted, rarely updated. When there are updates, they do not have any concurrency requirements.
+
+These are much simpler requirements than what general purpose distributed file system have to satisfy; these would include concurrent access to files, random access for reads and updates, posix compliance, and others. To satisfy those requirements, general DFSs are also pretty complex that are expensive to build and maintain.
+
+A different implementation of a distributed file system includes HDFS which is inspired by Google's GFS. This is one of the most widely used distributed file system that forms the main data storage platform for Hadoop. HDFS is primary aimed at processing very large data sets and distributes files across a cluster of commodity servers by splitting up files in fixed size chunks. HDFS is not particularly well suited for storing a very large number of relatively tiny files.
+
+### File Store
+
+It's possible to build a vastly simpler system for the class of applications that have simpler requirements as we have pointed out.
+
+* Large number of files but each file is relatively small
+* Access is limited to create, delete and get entire files
+* No updates to files that are already created (or it's feasible to delete the old file and create a new one)
+
+
+We call this system a Partitioned File Store (PFS) to distinguish it from other distributed file systems. This system needs to provide the following features:
+
+* CRD access to large number of small files
+* Scalability: Files should be distributed across a large number of commodity servers based on the storage requirement
+* Fault-tolerance: Each file should be replicated on multiple servers so that individual server failures do not reduce availability
+* Elasticity: It should be possible to add capacity to the cluster easily
+
+
+Apache Helix is a generic cluster management framework that makes it very easy to provide scalability, fault-tolerance and elasticity features.
+rsync can be easily used as a replication channel between servers so that each file gets replicated on multiple servers.
+
+### Design
+
+#### High Level
+
+* Partition the file system based on the file name
+* At any time a single writer can write, we call this a master
+* For redundancy, we need to have additional replicas called slave. Slaves can optionally serve reads
+* Slave replicates data from the master
+* When a master fails, a slave gets promoted to master
+
+#### Transaction Log
+
+Every write on the master will result in creation/deletion of one or more files. In order to maintain timeline consistency slaves need to apply the changes in the same order
+To facilitate this, the master logs each transaction in a file and each transaction is associated with an 64 bit ID in which the 32 LSB represents a sequence number and MSB represents the generation number
+The sequence number gets incremented on every transaction and the generation is incremented when a new master is elected
+
+#### Replication
+
+Replication is required for slaves to keep up with changes on the master. Every time the slave applies a change it checkpoints the last applied transaction ID.
+During restarts, this allows the slave to pull changes from the last checkpointed ID. Similar to master, the slave logs each transaction to the transaction logs but instead of generating new transaction ID, it uses the same ID generated by the master.
+
+
+#### Failover
+
+When a master fails, a new slave will be promoted to master. If the previous master node is reachable, then the new master will flush all the
+changes from previous the master before taking up mastership. The new master will record the end transaction ID of the current generation and then start a new generation
+with sequence starting from 1. After this the master will begin accepting writes.
+
+![Partitioned File Store](../images/PFS-Generic.png)
+
+
+
+### Rsync-based Solution
+
+![Rsync based File Store](../images/RSYNC_BASED_PFS.png)
+
+
+This application demonstrates a file store that uses rsync as the replication mechanism. One can envision a similar system where instead of using rsync, one
+can implement a custom solution to notify the slave of the changes and also provide an api to pull the change files.
+
+#### Concepts
+* file_store_dir: Root directory for the actual data files
+* change_log_dir: The transaction logs are generated under this folder
+* check_point_dir: The slave stores the check points ( last processed transaction) here
+
+#### Master
+* File server: This component supports file uploads and downloads and writes the files to ```file_store_dir```. This is not included in this application. The idea is that most applications have different ways of implementing this component and have some associated business logic. It is not hard to come up with such a component if needed.
+* File store watcher: This component watches the ```file_store_dir``` directory on the local file system for any changes and notifies the registered listeners of the changes
+* Change log generator: This registers as a listener of the file store watcher and on each notification logs the changes into a file under ```change_log_dir```
+
+#### Slave
+* File server: This component on the slave will only support reads
+* Cluster state observer: Slave observes the cluster state and is able to know who is the current master
+* Replicator: This has two subcomponents
+    - Periodic rsync of change log: This is a background process that periodically rsyncs the ```change_log_dir``` of the master to its local directory
+    - Change Log Watcher: This watches the ```change_log_dir``` for changes and notifies the registered listeners of the change
+    - On demand rsync invoker: This is registered as a listener to change log watcher and on every change invokes rsync to sync only the changed file
+
+#### Coordination
+
+The coordination between nodes is done by Helix. Helix does the partition management and assigns the partition to multiple nodes based on the replication factor. It elects one the nodes as master and designates others as slaves.
+It provides notifications to each node in the form of state transitions (Offline to Slave, Slave to Master). It also provides notifications when there is change is cluster state.
+This allows the slave to stop replicating from current master and start replicating from new master.
+
+In this application, we have only one partition but its very easy to extend it to support multiple partitions. By partitioning the file store, one can add new nodes and Helix will automatically
+re-distribute partitions among the nodes. To summarize, Helix provides partition management, fault tolerance and facilitates automated cluster expansion.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/recipes/service_discovery.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/recipes/service_discovery.md b/website/0.6.3/src/site/markdown/recipes/service_discovery.md
new file mode 100644
index 0000000..8ae2c85
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/recipes/service_discovery.md
@@ -0,0 +1,184 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+Service Discovery
+-----------------
+
+One of the common usage of ZooKeeper is to enable service discovery.
+The basic idea is that when a server starts up it advertises its configuration/metadata such as its hostname and port on ZooKeeper.
+This allows clients to dynamically discover the servers that are currently active. One can think of this like a service registry to which a server registers when it starts and
+is automatically deregistered when it shutdowns or crashes. In many cases it serves as an alternative to VIPs.
+
+The core idea behind this is to use ZooKeeper ephemeral nodes. The ephemeral nodes are created when the server registers and all its metadata is put into a ZNode.
+When the server shutdowns, ZooKeeper automatically removes this ZNode.
+
+There are two ways the clients can dynamically discover the active servers:
+
+### ZooKeeper Watch
+
+Clients can set a child watch under specific path on ZooKeeper.
+When a new service is registered/deregistered, ZooKeeper notifies the client via a watch event and the client can read the list of services. Even though this looks trivial,
+there are lot of things one needs to keep in mind like ensuring that you first set the watch back on ZooKeeper before reading data.
+
+
+### Poll
+
+Another approach is for the client to periodically read the ZooKeeper path and get the list of services.
+
+Both approaches have pros and cons, for example setting a watch might trigger herd effect if there are large number of clients. This is problematic, especially when servers are starting up.
+But the advantage to setting watches is that clients are immediately notified of a change which is not true in case of polling.
+In some cases, having both watches and polls makes sense; watch allows one to get notifications as soon as possible while poll provides a safety net if a watch event is missed because of code bug or ZooKeeper fails to notify.
+
+### Other Developer Considerations
+* What happens when the ZooKeeper session expires? All the watches and ephemeral nodes previously added or created by this server are lost. One needs to add the watches again, recreate the ephemeral nodes, and so on.
+* Due to network issues or Java GC pauses session expiry might happen again and again; this phenomenon is known as flapping. It\'s important for the server to detect this and deregister itself.
+
+### Other Operational Considerations
+* What if the node is behaving badly? One might kill the server, but it will lose the ability to debug. It would be nice to have the ability to mark a server as disabled and clients know that a node is disabled and will not contact that node.
+
+### Configuration Ownership
+
+This is an important aspect that is often ignored in the initial stages of your development. Typically, the service discovery pattern means that servers start up with some configuration which it simply puts into ZooKeeper. While this works well in the beginning, configuration management becomes very difficult since the servers themselves are statically configured. Any change in server configuration implies restarting the server. Ideally, it will be nice to have the ability to change configuration dynamically without having to restart a server.
+
+Ideally you want a hybrid solution, a node starts with minimal configuration and gets the rest of configuration from ZooKeeper.
+
+### Using Helix for Service Discovery
+
+Even though Helix has a higher-level abstraction in terms of state machines, constraints and objectives, service discovery is one of things has been a prevalent use case from the start.
+The controller uses the exact mechanism we described above to discover when new servers join the cluster. We create these ZNodes under /CLUSTERNAME/LIVEINSTANCES.
+Since at any time there is only one controller, we use a ZK watch to track the liveness of a server.
+
+This recipe simply demonstrates how one can re-use that part for implementing service discovery. This demonstrates multiple modes of service discovery:
+
+* POLL: The client reads from zookeeper at regular intervals 30 seconds. Use this if you have 100's of clients
+* WATCH: The client sets up watcher and gets notified of the changes. Use this if you have 10's of clients
+* NONE: This does neither of the above, but reads directly from zookeeper when ever needed
+
+Helix provides these additional features compared to other implementations available elsewhere:
+
+* It has the concept of disabling a node which means that a badly behaving node can be disabled using the Helix admin API
+* It automatically detects if a node connects/disconnects from zookeeper repeatedly and disables the node
+* Configuration management
+    * Allows one to set configuration via the admin API at various granulaties like cluster, instance, resource, partition
+    * Configurations can be dynamically changed
+    * The server is notified when configurations change
+
+
+### Checkout and Build
+
+```
+git clone https://git-wip-us.apache.org/repos/asf/helix.git
+cd helix
+git checkout tags/helix-0.6.3
+mvn clean install package -DskipTests
+cd recipes/service-discovery/target/service-discovery-pkg/bin
+chmod +x *
+```
+
+### Start ZooKeeper
+
+```
+./start-standalone-zookeeper 2199
+```
+
+### Run the Demo
+
+```
+./service-discovery-demo.sh
+```
+
+### Output
+
+```
+START:Service discovery demo mode:WATCH
+	Registering service
+		host.x.y.z_12000
+		host.x.y.z_12001
+		host.x.y.z_12002
+		host.x.y.z_12003
+		host.x.y.z_12004
+	SERVICES AVAILABLE
+		SERVICENAME 	HOST 			PORT
+		myServiceName 	host.x.y.z 		12000
+		myServiceName 	host.x.y.z 		12001
+		myServiceName 	host.x.y.z 		12002
+		myServiceName 	host.x.y.z 		12003
+		myServiceName 	host.x.y.z 		12004
+	Deregistering service:
+		host.x.y.z_12002
+	SERVICES AVAILABLE
+		SERVICENAME 	HOST 			PORT
+		myServiceName 	host.x.y.z 		12000
+		myServiceName 	host.x.y.z 		12001
+		myServiceName 	host.x.y.z 		12003
+		myServiceName 	host.x.y.z 		12004
+	Registering service:host.x.y.z_12002
+END:Service discovery demo mode:WATCH
+=============================================
+START:Service discovery demo mode:POLL
+	Registering service
+		host.x.y.z_12000
+		host.x.y.z_12001
+		host.x.y.z_12002
+		host.x.y.z_12003
+		host.x.y.z_12004
+	SERVICES AVAILABLE
+		SERVICENAME 	HOST 			PORT
+		myServiceName 	host.x.y.z 		12000
+		myServiceName 	host.x.y.z 		12001
+		myServiceName 	host.x.y.z 		12002
+		myServiceName 	host.x.y.z 		12003
+		myServiceName 	host.x.y.z 		12004
+	Deregistering service:
+		host.x.y.z_12002
+	Sleeping for poll interval:30000
+	SERVICES AVAILABLE
+		SERVICENAME 	HOST 			PORT
+		myServiceName 	host.x.y.z 		12000
+		myServiceName 	host.x.y.z 		12001
+		myServiceName 	host.x.y.z 		12003
+		myServiceName 	host.x.y.z 		12004
+	Registering service:host.x.y.z_12002
+END:Service discovery demo mode:POLL
+=============================================
+START:Service discovery demo mode:NONE
+	Registering service
+		host.x.y.z_12000
+		host.x.y.z_12001
+		host.x.y.z_12002
+		host.x.y.z_12003
+		host.x.y.z_12004
+	SERVICES AVAILABLE
+		SERVICENAME 	HOST 			PORT
+		myServiceName 	host.x.y.z 		12000
+		myServiceName 	host.x.y.z 		12001
+		myServiceName 	host.x.y.z 		12002
+		myServiceName 	host.x.y.z 		12003
+		myServiceName 	host.x.y.z 		12004
+	Deregistering service:
+		host.x.y.z_12000
+	SERVICES AVAILABLE
+		SERVICENAME 	HOST 			PORT
+		myServiceName 	host.x.y.z 		12001
+		myServiceName 	host.x.y.z 		12002
+		myServiceName 	host.x.y.z 		12003
+		myServiceName 	host.x.y.z 		12004
+	Registering service:host.x.y.z_12000
+END:Service discovery demo mode:NONE
+=============================================
+```

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/recipes/task_dag_execution.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/recipes/task_dag_execution.md b/website/0.6.3/src/site/markdown/recipes/task_dag_execution.md
new file mode 100644
index 0000000..f9757ab
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/recipes/task_dag_execution.md
@@ -0,0 +1,199 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+Distributed Task Execution
+--------------------------
+
+This recipe is intended to demonstrate how task dependencies can be modeled using primitives provided by Helix. A given task can be run with the desired amount of parallelism and will start only when upstream dependencies are met. The demo executes the task DAG described below using 10 workers. Although the demo starts the workers as threads, there is no requirement that all the workers need to run in the same process. In reality, these workers run on many different boxes on a cluster.  When worker fails, Helix takes care of re-assigning a failed task partition to a new worker.
+
+Redis is used as a result store. Any other suitable implementation for TaskResultStore can be plugged in.
+
+### Workflow
+
+#### Input
+
+10000 impression events and around 100 click events are pre-populated in task result store (redis).
+
+* **ImpEvent**: format: id,isFraudulent,country,gender
+
+* **ClickEvent**: format: id,isFraudulent,impEventId
+
+#### Stages
+
++ **FilterImps**: Filters impression where isFraudulent=true.
+
++ **FilterClicks**: Filters clicks where isFraudulent=true
+
++ **impCountsByGender**: Generates impression counts grouped by gender. It does this by incrementing the count for 'impression_gender_counts:<gender_value>' in the task result store (redis hash). Depends on: **FilterImps**
+
++ **impCountsByCountry**: Generates impression counts grouped by country. It does this by incrementing the count for 'impression_country_counts:<country_value>' in the task result store (redis hash). Depends on: **FilterClicks**
+
++ **impClickJoin**: Joins clicks with corresponding impression event using impEventId as the join key. Join is needed to pull dimensions not present in click event. Depends on: **FilterImps, FilterClicks**
+
++ **clickCountsByGender**: Generates click counts grouped by gender. It does this by incrementing the count for click_gender_counts:<gender_value> in the task result store (redis hash). Depends on: **impClickJoin**
+
++ **clickCountsByGender**: Generates click counts grouped by country. It does this by incrementing the count for click_country_counts:<country_value> in the task result store (redis hash). Depends on: **impClickJoin**
+
++ **report**: Reads from all aggregates generated by previous stages and prints them. Depends on: **impCountsByGender, impCountsByCountry, clickCountsByGender,clickCountsByGender**
+
+
+### Creating a DAG
+
+Each stage is represented as a Node along with the upstream dependency and desired parallelism.  Each stage is modeled as a resource in Helix using OnlineOffline state model. As part of an Offline to Online transition, we watch the external view of upstream resources and wait for them to transition to the online state. See Task.java for additional info.
+
+```
+Dag dag = new Dag();
+dag.addNode(new Node("filterImps", 10, ""));
+dag.addNode(new Node("filterClicks", 5, ""));
+dag.addNode(new Node("impClickJoin", 10, "filterImps,filterClicks"));
+dag.addNode(new Node("impCountsByGender", 10, "filterImps"));
+dag.addNode(new Node("impCountsByCountry", 10, "filterImps"));
+dag.addNode(new Node("clickCountsByGender", 5, "impClickJoin"));
+dag.addNode(new Node("clickCountsByCountry", 5, "impClickJoin"));
+dag.addNode(new Node("report",1,"impCountsByGender,impCountsByCountry,clickCountsByGender,clickCountsByCountry"));
+```
+
+### Demo
+
+In order to run the demo, use the following steps
+
+See http://redis.io/topics/quickstart on how to install redis server
+
+```
+Start redis e.g:
+./redis-server --port 6379
+
+git clone https://git-wip-us.apache.org/repos/asf/helix.git
+cd helix
+git checkout helix-0.6.3
+cd recipes/task-execution
+mvn clean install package -DskipTests
+cd target/task-execution-pkg/bin
+chmod +x task-execution-demo.sh
+./task-execution-demo.sh 2181 localhost 6379
+
+```
+
+Here\'s a visual representation of the DAG.
+
+```
+
+
+
+
+
+                       +-----------------+       +----------------+
+                       |   filterImps    |       |  filterClicks  |
+                       | (parallelism=10)|       | (parallelism=5)|
+                       +----------+-----++       +-------+--------+
+                       |          |     |                |
+                       |          |     |                |
+                       |          |     |                |
+                       |          |     +------->--------v------------+
+      +--------------<-+   +------v-------+    |  impClickJoin        |
+      |impCountsByGender   |impCountsByCountry | (parallelism=10)     |
+      |(parallelism=10)    |(parallelism=10)   ++-------------------+-+
+      +-----------+--+     +---+----------+     |                   |
+                  |            |                |                   |
+                  |            |                |                   |
+                  |            |       +--------v---------+       +-v-------------------+
+                  |            |       |clickCountsByGender       |clickCountsByCountry |
+                  |            |       |(parallelism=5)   |       |(parallelism=5)      |
+                  |            |       +----+-------------+       +---------------------+
+                  |            |            |                     |
+                  |            |            |                     |
+                  |            |            |                     |
+                  +----->+-----+>-----------v----+<---------------+
+                         | report                |
+                         |(parallelism=1)        |
+                         +-----------------------+
+
+```
+
+(credit for above ascii art: http://www.asciiflow.com)
+
+#### Output
+
+```
+Done populating dummy data
+Executing filter task for filterImps_3 for impressions_demo
+Executing filter task for filterImps_2 for impressions_demo
+Executing filter task for filterImps_0 for impressions_demo
+Executing filter task for filterImps_1 for impressions_demo
+Executing filter task for filterImps_4 for impressions_demo
+Executing filter task for filterClicks_3 for clicks_demo
+Executing filter task for filterClicks_1 for clicks_demo
+Executing filter task for filterImps_8 for impressions_demo
+Executing filter task for filterImps_6 for impressions_demo
+Executing filter task for filterClicks_2 for clicks_demo
+Executing filter task for filterClicks_0 for clicks_demo
+Executing filter task for filterImps_7 for impressions_demo
+Executing filter task for filterImps_5 for impressions_demo
+Executing filter task for filterClicks_4 for clicks_demo
+Executing filter task for filterImps_9 for impressions_demo
+Running AggTask for impCountsByGender_3 for filtered_impressions_demo gender
+Running AggTask for impCountsByGender_2 for filtered_impressions_demo gender
+Running AggTask for impCountsByGender_0 for filtered_impressions_demo gender
+Running AggTask for impCountsByGender_9 for filtered_impressions_demo gender
+Running AggTask for impCountsByGender_1 for filtered_impressions_demo gender
+Running AggTask for impCountsByGender_4 for filtered_impressions_demo gender
+Running AggTask for impCountsByCountry_4 for filtered_impressions_demo country
+Running AggTask for impCountsByGender_5 for filtered_impressions_demo gender
+Executing JoinTask for impClickJoin_2
+Running AggTask for impCountsByCountry_3 for filtered_impressions_demo country
+Running AggTask for impCountsByCountry_1 for filtered_impressions_demo country
+Running AggTask for impCountsByCountry_0 for filtered_impressions_demo country
+Running AggTask for impCountsByCountry_2 for filtered_impressions_demo country
+Running AggTask for impCountsByGender_6 for filtered_impressions_demo gender
+Executing JoinTask for impClickJoin_1
+Executing JoinTask for impClickJoin_0
+Executing JoinTask for impClickJoin_3
+Running AggTask for impCountsByGender_8 for filtered_impressions_demo gender
+Executing JoinTask for impClickJoin_4
+Running AggTask for impCountsByGender_7 for filtered_impressions_demo gender
+Running AggTask for impCountsByCountry_5 for filtered_impressions_demo country
+Running AggTask for impCountsByCountry_6 for filtered_impressions_demo country
+Executing JoinTask for impClickJoin_9
+Running AggTask for impCountsByCountry_8 for filtered_impressions_demo country
+Running AggTask for impCountsByCountry_7 for filtered_impressions_demo country
+Executing JoinTask for impClickJoin_5
+Executing JoinTask for impClickJoin_6
+Running AggTask for impCountsByCountry_9 for filtered_impressions_demo country
+Executing JoinTask for impClickJoin_8
+Executing JoinTask for impClickJoin_7
+Running AggTask for clickCountsByCountry_1 for joined_clicks_demo country
+Running AggTask for clickCountsByCountry_0 for joined_clicks_demo country
+Running AggTask for clickCountsByCountry_2 for joined_clicks_demo country
+Running AggTask for clickCountsByCountry_3 for joined_clicks_demo country
+Running AggTask for clickCountsByGender_1 for joined_clicks_demo gender
+Running AggTask for clickCountsByCountry_4 for joined_clicks_demo country
+Running AggTask for clickCountsByGender_3 for joined_clicks_demo gender
+Running AggTask for clickCountsByGender_2 for joined_clicks_demo gender
+Running AggTask for clickCountsByGender_4 for joined_clicks_demo gender
+Running AggTask for clickCountsByGender_0 for joined_clicks_demo gender
+Running reports task
+Impression counts per country
+{CANADA=1940, US=1958, CHINA=2014, UNKNOWN=2022, UK=1946}
+Click counts per country
+{US=24, CANADA=14, CHINA=26, UNKNOWN=14, UK=22}
+Impression counts per gender
+{F=3325, UNKNOWN=3259, M=3296}
+Click counts per gender
+{F=33, UNKNOWN=32, M=35}
+```

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_admin.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_admin.md b/website/0.6.3/src/site/markdown/tutorial_admin.md
new file mode 100644
index 0000000..5c4b1d5
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_admin.md
@@ -0,0 +1,408 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Admin Operations</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Admin Operations
+
+Helix provides a set of admin APIs for cluster management operations. They are supported via:
+
+* Java API
+* Command Line Interface
+* REST Interface via helix-admin-webapp
+
+### Java API
+See interface [_org.apache.helix.HelixAdmin_](http://helix.apache.org/javadocs/0.6.3/reference/org/apache/helix/HelixAdmin.html)
+
+### Command Line Interface
+The command line tool comes with helix-core package:
+
+Get the command line tool:
+
+```
+git clone https://git-wip-us.apache.org/repos/asf/helix.git
+cd helix
+git checkout tags/helix-0.6.3
+./build
+cd helix-core/target/helix-core-pkg/bin
+chmod +x *.sh
+```
+
+Get help:
+
+```
+./helix-admin.sh --help
+```
+
+All other commands have this form:
+
+```
+./helix-admin.sh --zkSvr <ZookeeperServerAddress> <command> <parameters>
+```
+
+#### Supported Commands
+
+| Command Syntax | Description |
+| -------------- | ----------- |
+| _\-\-activateCluster \<clusterName controllerCluster true/false\>_ | Enable/disable a cluster in distributed controller mode |
+| _\-\-addCluster \<clusterName\>_ | Add a new cluster |
+| _\-\-addIdealState \<clusterName resourceName fileName.json\>_ | Add an ideal state to a cluster |
+| _\-\-addInstanceTag \<clusterName instanceName tag\>_ | Add a tag to an instance |
+| _\-\-addNode \<clusterName instanceId\>_ | Add an instance to a cluster |
+| _\-\-addResource \<clusterName resourceName partitionNumber stateModelName\>_ | Add a new resource to a cluster |
+| _\-\-addResourceProperty \<clusterName resourceName propertyName propertyValue\>_ | Add a resource property |
+| _\-\-addStateModelDef \<clusterName fileName.json\>_ | Add a State model definition to a cluster |
+| _\-\-dropCluster \<clusterName\>_ | Delete a cluster |
+| _\-\-dropNode \<clusterName instanceId\>_ | Remove a node from a cluster |
+| _\-\-dropResource \<clusterName resourceName\>_ | Remove an existing resource from a cluster |
+| _\-\-enableCluster \<clusterName true/false\>_ | Enable/disable a cluster |
+| _\-\-enableInstance \<clusterName instanceId true/false\>_ | Enable/disable an instance |
+| _\-\-enablePartition \<true/false clusterName nodeId resourceName partitionName\>_ | Enable/disable a partition |
+| _\-\-getConfig \<configScope configScopeArgs configKeys\>_ | Get user configs |
+| _\-\-getConstraints \<clusterName constraintType\>_ | Get constraints |
+| _\-\-help_ | print help information |
+| _\-\-instanceGroupTag \<instanceTag\>_ | Specify instance group tag, used with rebalance command |
+| _\-\-listClusterInfo \<clusterName\>_ | Show information of a cluster |
+| _\-\-listClusters_ | List all clusters |
+| _\-\-listInstanceInfo \<clusterName instanceId\>_ | Show information of an instance |
+| _\-\-listInstances \<clusterName\>_ | List all instances in a cluster |
+| _\-\-listPartitionInfo \<clusterName resourceName partitionName\>_ | Show information of a partition |
+| _\-\-listResourceInfo \<clusterName resourceName\>_ | Show information of a resource |
+| _\-\-listResources \<clusterName\>_ | List all resources in a cluster |
+| _\-\-listStateModel \<clusterName stateModelName\>_ | Show information of a state model |
+| _\-\-listStateModels \<clusterName\>_ | List all state models in a cluster |
+| _\-\-maxPartitionsPerNode \<maxPartitionsPerNode\>_ | Specify the max partitions per instance, used with addResourceGroup command |
+| _\-\-rebalance \<clusterName resourceName replicas\>_ | Rebalance a resource |
+| _\-\-removeConfig \<configScope configScopeArgs configKeys\>_ | Remove user configs |
+| _\-\-removeConstraint \<clusterName constraintType constraintId\>_ | Remove a constraint |
+| _\-\-removeInstanceTag \<clusterName instanceId tag\>_ | Remove a tag from an instance |
+| _\-\-removeResourceProperty \<clusterName resourceName propertyName\>_ | Remove a resource property |
+| _\-\-resetInstance \<clusterName instanceId\>_ | Reset all erroneous partitions on an instance |
+| _\-\-resetPartition \<clusterName instanceId resourceName partitionName\>_ | Reset an erroneous partition |
+| _\-\-resetResource \<clusterName resourceName\>_ | Reset all erroneous partitions of a resource |
+| _\-\-setConfig \<configScope configScopeArgs configKeyValueMap\>_ | Set user configs |
+| _\-\-setConstraint \<clusterName constraintType constraintId constraintKeyValueMap\>_ | Set a constraint |
+| _\-\-swapInstance \<clusterName oldInstance newInstance\>_ | Swap an old instance with a new instance |
+| _\-\-zkSvr \<ZookeeperServerAddress\>_ | Provide zookeeper address |
+
+### REST Interface
+
+The REST interface comes wit helix-admin-webapp package:
+
+```
+git clone https://git-wip-us.apache.org/repos/asf/helix.git
+cd helix
+git checkout tags/helix-0.6.3
+./build
+cd helix-admin-webapp/target/helix-admin-webapp-pkg/bin
+chmod +x *.sh
+./run-rest-admin.sh --zkSvr <zookeeperAddress> --port <port> // make sure ZooKeeper is running
+```
+
+#### URL and support methods
+
+* _/clusters_
+    * List all clusters
+
+    ```
+    curl http://localhost:8100/clusters
+    ```
+
+    * Add a cluster
+
+    ```
+    curl -d 'jsonParameters={"command":"addCluster","clusterName":"MyCluster"}' -H "Content-Type: application/json" http://localhost:8100/clusters
+    ```
+
+* _/clusters/{clusterName}_
+    * List cluster information
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster
+    ```
+
+    * Enable/disable a cluster in distributed controller mode
+
+    ```
+    curl -d 'jsonParameters={"command":"activateCluster","grandCluster":"MyControllerCluster","enabled":"true"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster
+    ```
+
+    * Remove a cluster
+
+    ```
+    curl -X DELETE http://localhost:8100/clusters/MyCluster
+    ```
+
+* _/clusters/{clusterName}/resourceGroups_
+    * List all resources in a cluster
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/resourceGroups
+    ```
+
+    * Add a resource to cluster
+
+    ```
+    curl -d 'jsonParameters={"command":"addResource","resourceGroupName":"MyDB","partitions":"8","stateModelDefRef":"MasterSlave" }' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/resourceGroups
+    ```
+
+* _/clusters/{clusterName}/resourceGroups/{resourceName}_
+    * List resource information
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB
+    ```
+
+    * Drop a resource
+
+    ```
+    curl -X DELETE http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB
+    ```
+
+    * Reset all erroneous partitions of a resource
+
+    ```
+    curl -d 'jsonParameters={"command":"resetResource"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB
+    ```
+
+* _/clusters/{clusterName}/resourceGroups/{resourceName}/idealState_
+    * Rebalance a resource
+
+    ```
+    curl -d 'jsonParameters={"command":"rebalance","replicas":"3"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB/idealState
+    ```
+
+    * Add an ideal state
+
+    ```
+    echo jsonParameters={
+    "command":"addIdealState"
+       }&newIdealState={
+      "id" : "MyDB",
+      "simpleFields" : {
+        "IDEAL_STATE_MODE" : "AUTO",
+        "NUM_PARTITIONS" : "8",
+        "REBALANCE_MODE" : "SEMI_AUTO",
+        "REPLICAS" : "0",
+        "STATE_MODEL_DEF_REF" : "MasterSlave",
+        "STATE_MODEL_FACTORY_NAME" : "DEFAULT"
+      },
+      "listFields" : {
+      },
+      "mapFields" : {
+        "MyDB_0" : {
+          "localhost_1001" : "MASTER",
+          "localhost_1002" : "SLAVE"
+        }
+      }
+    }
+    > newIdealState.json
+    curl -d @'./newIdealState.json' -H 'Content-Type: application/json' http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB/idealState
+    ```
+
+    * Add resource property
+
+    ```
+    curl -d 'jsonParameters={"command":"addResourceProperty","REBALANCE_TIMER_PERIOD":"500"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB/idealState
+    ```
+
+* _/clusters/{clusterName}/resourceGroups/{resourceName}/externalView_
+    * Show resource external view
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/resourceGroups/MyDB/externalView
+    ```
+* _/clusters/{clusterName}/instances_
+    * List all instances
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/instances
+    ```
+
+    * Add an instance
+
+    ```
+    curl -d 'jsonParameters={"command":"addInstance","instanceNames":"localhost_1001"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/instances
+    ```
+
+    * Swap an instance
+
+    ```
+    curl -d 'jsonParameters={"command":"swapInstance","oldInstance":"localhost_1001", "newInstance":"localhost_1002"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/instances
+    ```
+* _/clusters/{clusterName}/instances/{instanceName}_
+    * Show instance information
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/instances/localhost_1001
+    ```
+
+    * Enable/disable an instance
+
+    ```
+    curl -d 'jsonParameters={"command":"enableInstance","enabled":"false"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/instances/localhost_1001
+    ```
+
+    * Drop an instance
+
+    ```
+    curl -X DELETE http://localhost:8100/clusters/MyCluster/instances/localhost_1001
+    ```
+
+    * Disable/enable partitions on an instance
+
+    ```
+    curl -d 'jsonParameters={"command":"enablePartition","resource": "MyDB","partition":"MyDB_0",  "enabled" : "false"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/instances/localhost_1001
+    ```
+
+    * Reset an erroneous partition on an instance
+
+    ```
+    curl -d 'jsonParameters={"command":"resetPartition","resource": "MyDB","partition":"MyDB_0"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/instances/localhost_1001
+    ```
+
+    * Reset all erroneous partitions on an instance
+
+    ```
+    curl -d 'jsonParameters={"command":"resetInstance"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/instances/localhost_1001
+    ```
+
+* _/clusters/{clusterName}/configs_
+    * Get user cluster level config
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/configs/cluster
+    ```
+
+    * Set user cluster level config
+
+    ```
+    curl -d 'jsonParameters={"command":"setConfig","configs":"key1=value1,key2=value2"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/configs/cluster
+    ```
+
+    * Remove user cluster level config
+
+    ```
+    curl -d 'jsonParameters={"command":"removeConfig","configs":"key1,key2"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/configs/cluster
+    ```
+
+    * Get/set/remove user participant level config
+
+    ```
+    curl -d 'jsonParameters={"command":"setConfig","configs":"key1=value1,key2=value2"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/configs/participant/localhost_1001
+    ```
+
+    * Get/set/remove resource level config
+
+    ```
+    curl -d 'jsonParameters={"command":"setConfig","configs":"key1=value1,key2=value2"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/configs/resource/MyDB
+    ```
+
+* _/clusters/{clusterName}/controller_
+    * Show controller information
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/Controller
+    ```
+
+    * Enable/disable cluster
+
+    ```
+    curl -d 'jsonParameters={"command":"enableCluster","enabled":"false"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/Controller
+    ```
+
+* _/zkPath/{path}_
+    * Get information for zookeeper path
+
+    ```
+    curl http://localhost:8100/zkPath/MyCluster
+    ```
+
+* _/clusters/{clusterName}/StateModelDefs_
+    * Show all state model definitions
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/StateModelDefs
+    ```
+
+    * Add a state mdoel definition
+
+    ```
+    echo jsonParameters={
+      "command":"addStateModelDef"
+    }&newStateModelDef={
+      "id" : "OnlineOffline",
+      "simpleFields" : {
+        "INITIAL_STATE" : "OFFLINE"
+      },
+      "listFields" : {
+        "STATE_PRIORITY_LIST" : [ "ONLINE", "OFFLINE", "DROPPED" ],
+        "STATE_TRANSITION_PRIORITYLIST" : [ "OFFLINE-ONLINE", "ONLINE-OFFLINE", "OFFLINE-DROPPED" ]
+      },
+      "mapFields" : {
+        "DROPPED.meta" : {
+          "count" : "-1"
+        },
+        "OFFLINE.meta" : {
+          "count" : "-1"
+        },
+        "OFFLINE.next" : {
+          "DROPPED" : "DROPPED",
+          "ONLINE" : "ONLINE"
+        },
+        "ONLINE.meta" : {
+          "count" : "R"
+        },
+        "ONLINE.next" : {
+          "DROPPED" : "OFFLINE",
+          "OFFLINE" : "OFFLINE"
+        }
+      }
+    }
+    > newStateModelDef.json
+    curl -d @'./untitled.txt' -H 'Content-Type: application/json' http://localhost:8100/clusters/MyCluster/StateModelDefs
+    ```
+
+* _/clusters/{clusterName}/StateModelDefs/{stateModelDefName}_
+    * Show a state model definition
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/StateModelDefs/OnlineOffline
+    ```
+
+* _/clusters/{clusterName}/constraints/{constraintType}_
+    * Show all contraints
+
+    ```
+    curl http://localhost:8100/clusters/MyCluster/constraints/MESSAGE_CONSTRAINT
+    ```
+
+    * Set a contraint
+
+    ```
+    curl -d 'jsonParameters={"constraintAttributes":"RESOURCE=MyDB,CONSTRAINT_VALUE=1"}' -H "Content-Type: application/json" http://localhost:8100/clusters/MyCluster/constraints/MESSAGE_CONSTRAINT/MyConstraint
+    ```
+
+    * Remove a constraint
+
+    ```
+    curl -X DELETE http://localhost:8100/clusters/MyCluster/constraints/MESSAGE_CONSTRAINT/MyConstraint
+    ```

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_agent.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_agent.md b/website/0.6.3/src/site/markdown/tutorial_agent.md
new file mode 100644
index 0000000..235d71d
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_agent.md
@@ -0,0 +1,169 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Helix Agent</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Helix Agent (for non-JVM systems)
+
+Not every distributed system is written on the JVM, but many systems would benefit from the cluster management features that Helix provides. To make a non-JVM system work with Helix, you can use the Helix Agent module.
+
+### What is Helix Agent?
+
+Helix is built on the following assumption: if your distributed resource is modeled by a finite state machine, then Helix can tell participants when they should transition between states. In the Java API, this means implementing transition callbacks. In the Helix agent API, this means providing commands than can run for each transition.
+
+These commands could do anything behind the scenes; Helix only requires that they exit once the state transition is complete.
+
+### Configuring Transition Commands
+
+Here's how to tell Helix which commands to run on state transitions:
+
+#### Java
+
+Using the Java API, first get a configuration scope (the Helix agent supports both cluster and resource scopes, picking resource first if it is available):
+
+```
+// Cluster scope
+HelixConfigScope scope =
+    new HelixConfigScopeBuilder(ConfigScopeProperty.CLUSTER).forCluster(clusterName).build();
+
+// Resource scope
+HelixConfigScope scope =
+    new HelixConfigScopeBuilder(ConfigScopeProperty.RESOURCE).forCluster(clusterName).forResource(resourceName).build();
+```
+
+Then, specify the command to run for each state transition:
+
+```
+// Get the configuration accessor
+ConfigAccessor configAccessor = new ConfigAccessor(_gZkClient);
+
+// Specify the script for OFFLINE --> ONLINE
+CommandConfig.Builder builder = new CommandConfig.Builder();
+CommandConfig cmdConfig =
+    builder.setTransition("OFFLINE", "ONLINE").setCommand("simpleHttpClient.py OFFLINE-ONLINE")
+        .setCommandWorkingDir(workingDir)
+        .setCommandTimeout("5000L") // optional: ms to wait before failing
+        .setPidFile(pidFile) // optional: for daemon-like systems that will write the process id to a file
+        .build();
+configAccessor.set(scope, cmdConfig.toKeyValueMap());
+
+// Specify the script for ONLINE --> OFFLINE
+builder = new CommandConfig.Builder();
+cmdConfig =
+    builder.setTransition("ONLINE", "OFFLINE").setCommand("simpleHttpClient.py ONLINE-OFFLINE")
+        .setCommandWorkingDir(workingDir)
+        .build();
+configAccessor.set(scope, cmdConfig.toKeyValueMap());
+
+// Specify NOP for OFFLINE --> DROPPED
+builder = new CommandConfig.Builder();
+cmdConfig =
+    builder.setTransition("OFFLINE", "DROPPED")
+        .setCommand(CommandAttribute.NOP.getName())
+        .build();
+configAccessor.set(scope, cmdConfig.toKeyValueMap());
+```
+
+In this example, we have a program called simpleHttpClient.py that we call for all transitions, only changing the arguments that are passed in. However, there is no requirement that each transition invoke the same program; this API allows running arbitrary commands in arbitrary directories with arbitrary arguments.
+
+Notice that that for the OFFLINE \-\-\> DROPPED transition, we do not run any command (specifically, we specify the NOP command). This just tells Helix that the system doesn't care about when things are dropped, and it can consider the transition already done.
+
+#### Command Line
+
+It is also possible to configure everything directly from the command line. Here's how that would look for cluster-wide configuration:
+
+```
+# Specify the script for OFFLINE --> ONLINE
+/helix-admin.sh --zkSvr localhost:2181 --setConfig CLUSTER clusterName OFFLINE-ONLINE.command="simpleHttpClient.py OFFLINE-ONLINE",OFFLINE-ONLINE.workingDir="/path/to/script", OFFLINE-ONLINE.command.pidfile="/path/to/pidfile"
+
+# Specify the script for ONLINE --> OFFLINE
+/helix-admin.sh --zkSvr localhost:2181 --setConfig CLUSTER clusterName ONLINE-OFFLINE.command="simpleHttpClient.py ONLINE-OFFLINE",ONLINE-OFFLINE.workingDir="/path/to/script", OFFLINE-ONLINE.command.pidfile="/path/to/pidfile"
+
+# Specify NOP for OFFLINE --> DROPPED
+/helix-admin.sh --zkSvr localhost:2181 --setConfig CLUSTER clusterName ONLINE-OFFLINE.command="nop"
+```
+
+Like in the Java configuration, it is also possible to specify a resource scope instead of a cluster scope:
+
+```
+# Specify the script for OFFLINE --> ONLINE
+/helix-admin.sh --zkSvr localhost:2181 --setConfig RESOURCE clusterName,resourceName OFFLINE-ONLINE.command="simpleHttpClient.py OFFLINE-ONLINE",OFFLINE-ONLINE.workingDir="/path/to/script", OFFLINE-ONLINE.command.pidfile="/path/to/pidfile"
+```
+
+### Starting the Agent
+
+There should be an agent running for every participant you have running. Ideally, its lifecycle should match that of the participant. Here, we have a simple long-running participant called simpleHttpServer.py. Its only purpose is to record state transitions.
+
+Here are some ways that you can start the Helix agent:
+
+#### Java
+
+```
+// Start your application process
+ExternalCommand serverCmd = ExternalCommand.start(workingDir + "/simpleHttpServer.py");
+
+// Start the agent
+Thread agentThread = new Thread() {
+  @Override
+  public void run() {
+    while(!isInterrupted()) {
+      try {
+        HelixAgentMain.main(new String[] {
+            "--zkSvr", zkAddr, "--cluster", clusterName, "--instanceName", instanceName,
+            "--stateModel", "OnlineOffline"
+        });
+      } catch (InterruptedException e) {
+        LOG.info("Agent thread interrupted", e);
+        interrupt();
+      } catch (Exception e) {
+        LOG.error("Exception start helix-agent", e);
+      }
+    }
+  }
+};
+agentThread.start();
+
+// Wait for the process to terminate (either intentionally or unintentionally)
+serverCmd.waitFor();
+
+// Kill the agent
+agentThread.interrupt();
+```
+
+#### Command Line
+
+```
+# Build Helix and start the agent
+mvn clean install -DskipTests
+chmod +x helix-agent/target/helix-agent-pkg/bin/*
+helix-agent/target/helix-agent-pkg/bin/start-helix-agent.sh --zkSvr zkAddr1,zkAddr2 --cluster clusterName --instanceName instanceName --stateModel OnlineOffline
+
+# Here, you can define your own logic to terminate this agent when your process terminates
+...
+```
+
+### Example
+
+[Here](https://git-wip-us.apache.org/repos/asf?p=helix.git;a=blob;f=helix-agent/src/test/java/org/apache/helix/agent/TestHelixAgent.java;h=ccf64ce5544207c7e48261682ea69945b71da7f1;hb=refs/heads/master) is a basic system that uses the Helix agent package.
+
+### Notes
+
+As you may have noticed from the examples, the participant program and the state transition program are two different programs. The former is a _long-running_ process that is directly tied to the Helix agent. The latter is a process that only exists while a state transition is underway. Despite this, these two processes should be intertwined. The transition command will need to communicate to the participant to actually complete the state transition and the participant will need to communicate whether or not this was successful. The implementation of this protocol is the responsibility of the system.
\ No newline at end of file

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_controller.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_controller.md b/website/0.6.3/src/site/markdown/tutorial_controller.md
new file mode 100644
index 0000000..ad782e5
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_controller.md
@@ -0,0 +1,159 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Controller</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Controller
+
+Next, let\'s implement the controller.  This is the brain of the cluster.  Helix makes sure there is exactly one active controller running the cluster.
+
+### Start a Connection
+
+The Helix manager requires the following parameters:
+
+* clusterName: A logical name to represent the group of nodes
+* instanceName: A logical name of the process creating the manager instance. Generally this is host:port
+* instanceType: Type of the process. This can be one of the following types, in this case use CONTROLLER:
+    * CONTROLLER: Process that controls the cluster, any number of controllers can be started but only one will be active at any given time
+    * PARTICIPANT: Process that performs the actual task in the distributed system
+    * SPECTATOR: Process that observes the changes in the cluster
+    * ADMIN: To carry out system admin actions
+* zkConnectString: Connection string to ZooKeeper. This is of the form host1:port1,host2:port2,host3:port3
+
+```
+manager = HelixManagerFactory.getZKHelixManager(clusterName,
+                                                instanceName,
+                                                instanceType,
+                                                zkConnectString);
+```
+
+### Controller Code
+
+The Controller needs to know about all changes in the cluster. Helix takes care of this with the default implementation.
+If you need additional functionality, see GenericHelixController on how to configure the pipeline.
+
+```
+manager = HelixManagerFactory.getZKHelixManager(clusterName,
+                                                instanceName,
+                                                InstanceType.CONTROLLER,
+                                                zkConnectString);
+manager.connect();
+GenericHelixController controller = new GenericHelixController();
+manager.addConfigChangeListener(controller);
+manager.addLiveInstanceChangeListener(controller);
+manager.addIdealStateChangeListener(controller);
+manager.addExternalViewChangeListener(controller);
+manager.addControllerListener(controller);
+```
+The snippet above shows how the controller is started. You can also start the controller using command line interface.
+
+```
+cd helix/helix-core/target/helix-core-pkg/bin
+./run-helix-controller.sh --zkSvr <Zookeeper ServerAddress (Required)>  --cluster <Cluster name (Required)>
+```
+
+### Controller Deployment Modes
+
+Helix provides multiple options to deploy the controller.
+
+#### STANDALONE
+
+The Controller can be started as a separate process to manage a cluster. This is the recommended approach. However, since one controller can be a single point of failure, multiple controller processes are required for reliability.  Even if multiple controllers are running, only one will be actively managing the cluster at any time and is decided by a leader-election process. If the leader fails, another leader will take over managing the cluster.
+
+Even though we recommend this method of deployment, it has the drawback of having to manage an additional service for each cluster. See the Controller as a Service option.
+
+#### EMBEDDED
+
+If setting up a separate controller process is not viable, then it is possible to embed the controller as a library in each of the participants.
+
+#### CONTROLLER AS A SERVICE
+
+One of the cool features we added in Helix was to use a set of controllers to manage a large number of clusters.
+
+For example if you have X clusters to be managed, instead of deploying X*3 (3 controllers for fault tolerance) controllers for each cluster, one can deploy just 3 controllers.  Each controller can manage X/3 clusters.  If any controller fails, the remaining two will manage X/2 clusters.
+
+Next, let\'s implement the controller.  This is the brain of the cluster.  Helix makes sure there is exactly one active controller running the cluster.
+
+### Start the Helix agent
+
+
+It requires the following parameters:
+
+* clusterName: A logical name to represent the group of nodes
+* instanceName: A logical name of the process creating the manager instance. Generally this is host:port.
+* instanceType: Type of the process. This can be one of the following types, in this case use CONTROLLER:
+    * CONTROLLER: Process that controls the cluster, any number of controllers can be started but only one will be active at any given time.
+    * PARTICIPANT: Process that performs the actual task in the distributed system.
+    * SPECTATOR: Process that observes the changes in the cluster.
+    * ADMIN: To carry out system admin actions.
+* zkConnectString: Connection string to Zookeeper. This is of the form host1:port1,host2:port2,host3:port3.
+
+```
+      manager = HelixManagerFactory.getZKHelixManager(clusterName,
+                                                      instanceName,
+                                                      instanceType,
+                                                      zkConnectString);
+```
+
+### Controller Code
+
+The Controller needs to know about all changes in the cluster. Helix takes care of this with the default implementation.
+If you need additional functionality, see GenericHelixController on how to configure the pipeline.
+
+```
+      manager = HelixManagerFactory.getZKHelixManager(clusterName,
+                                                          instanceName,
+                                                          InstanceType.CONTROLLER,
+                                                          zkConnectString);
+     manager.connect();
+     GenericHelixController controller = new GenericHelixController();
+     manager.addConfigChangeListener(controller);
+     manager.addLiveInstanceChangeListener(controller);
+     manager.addIdealStateChangeListener(controller);
+     manager.addExternalViewChangeListener(controller);
+     manager.addControllerListener(controller);
+```
+The snippet above shows how the controller is started. You can also start the controller using command line interface.
+
+```
+cd helix/helix-core/target/helix-core-pkg/bin
+./run-helix-controller.sh --zkSvr <Zookeeper ServerAddress (Required)>  --cluster <Cluster name (Required)>
+```
+
+### Controller Deployment Modes
+
+Helix provides multiple options to deploy the controller.
+
+#### STANDALONE
+
+The Controller can be started as a separate process to manage a cluster. This is the recommended approach. However, since one controller can be a single point of failure, multiple controller processes are required for reliability.  Even if multiple controllers are running, only one will be actively managing the cluster at any time and is decided by a leader-election process. If the leader fails, another leader will take over managing the cluster.
+
+Even though we recommend this method of deployment, it has the drawback of having to manage an additional service for each cluster. See Controller As a Service option.
+
+#### EMBEDDED
+
+If setting up a separate controller process is not viable, then it is possible to embed the controller as a library in each of the participants.
+
+#### CONTROLLER AS A SERVICE
+
+One of the cool features we added in Helix is to use a set of controllers to manage a large number of clusters.
+
+For example if you have X clusters to be managed, instead of deploying X*3 (3 controllers for fault tolerance) controllers for each cluster, one can deploy just 3 controllers.  Each controller can manage X/3 clusters.  If any controller fails, the remaining two will manage X/2 clusters.

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_health.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_health.md b/website/0.6.3/src/site/markdown/tutorial_health.md
new file mode 100644
index 0000000..03b1dcc
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_health.md
@@ -0,0 +1,46 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Customizing Heath Checks</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Customizing Health Checks
+
+In this chapter, we\'ll learn how to customize health checks based on metrics of your distributed system.
+
+### Health Checks
+
+Note: _this in currently in development mode, not yet ready for production._
+
+Helix provides the ability for each node in the system to report health metrics on a periodic basis.
+
+Helix supports multiple ways to aggregate these metrics:
+
+* SUM
+* AVG
+* EXPONENTIAL DECAY
+* WINDOW
+
+Helix persists the aggregated value only.
+
+Applications can define a threshold on the aggregate values according to the SLAs, and when the SLA is violated Helix will fire an alert.
+Currently Helix only fires an alert, but in a future release we plan to use these metrics to either mark the node dead or load balance the partitions.
+This feature will be valuable for distributed systems that support multi-tenancy and have a large variation in work load patterns.  In addition, this can be used to detect skewed partitions (hotspots) and rebalance the cluster.
+

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_messaging.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_messaging.md b/website/0.6.3/src/site/markdown/tutorial_messaging.md
new file mode 100644
index 0000000..22a6a10
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_messaging.md
@@ -0,0 +1,70 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Messaging</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Messaging
+
+In this chapter, we\'ll learn about messaging, a convenient feature in Helix for sending messages between nodes of a cluster.  This is an interesting feature that is quite useful in practice. It is common that nodes in a distributed system require a mechanism to interact with each other.
+
+### Example: Bootstrapping a Replica
+
+Consider a search system  where the index replica starts up and it does not have an index. A typical solution is to get the index from a common location, or to copy the index from another replica.
+
+Helix provides a messaging API for intra-cluster communication between nodes in the system.  This API provides a mechanism to specify the message recipient in terms of resource, partition, and state rather than specifying hostnames.  Helix ensures that the message is delivered to all of the required recipients. In this particular use case, the instance can specify the recipient criteria as all replicas of the desired partition to bootstrap.
+Since Helix is aware of the global state of the system, it can send the message to the appropriate nodes. Once the nodes respond, Helix provides the bootstrapping replica with all the responses.
+
+This is a very generic API and can also be used to schedule various periodic tasks in the cluster, such as data backups, log cleanup, etc.
+System Admins can also perform ad-hoc tasks, such as on-demand backups or a system command (such as rm -rf ;) across all nodes of the cluster
+
+```
+ClusterMessagingService messagingService = manager.getMessagingService();
+
+// Construct the Message
+Message requestBackupUriRequest = new Message(
+    MessageType.USER_DEFINE_MSG, UUID.randomUUID().toString());
+requestBackupUriRequest
+    .setMsgSubType(BootstrapProcess.REQUEST_BOOTSTRAP_URL);
+requestBackupUriRequest.setMsgState(MessageState.NEW);
+
+// Set the Recipient criteria: all nodes that satisfy the criteria will receive the message
+Criteria recipientCriteria = new Criteria();
+recipientCriteria.setInstanceName("%");
+recipientCriteria.setRecipientInstanceType(InstanceType.PARTICIPANT);
+recipientCriteria.setResource("MyDB");
+recipientCriteria.setPartition("");
+
+// Should be processed only by process(es) that are active at the time of sending the message
+// This means if the recipient is restarted after message is sent, it will not be processe.
+recipientCriteria.setSessionSpecific(true);
+
+// wait for 30 seconds
+int timeout = 30000;
+
+// the handler that will be invoked when any recipient responds to the message.
+BootstrapReplyHandler responseHandler = new BootstrapReplyHandler();
+
+// this will return only after all recipients respond or after timeout
+int sentMessageCount = messagingService.sendAndWait(recipientCriteria,
+    requestBackupUriRequest, responseHandler, timeout);
+```
+
+See HelixManager.DefaultMessagingService in the [Javadocs](http://helix.apache.org/javadocs/0.6.3/reference/org/apache/helix/messaging/DefaultMessagingService.html) for more information.

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_participant.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_participant.md b/website/0.6.3/src/site/markdown/tutorial_participant.md
new file mode 100644
index 0000000..cb38e45
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_participant.md
@@ -0,0 +1,102 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Participant</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Participant
+
+In this chapter, we\'ll learn how to implement a __Participant__, which is a primary functional component of a distributed system.
+
+
+### Start a Connection
+
+The Helix manager is a common component that connects each system component with the controller.
+
+It requires the following parameters:
+
+* clusterName: A logical name to represent the group of nodes
+* instanceName: A logical name of the process creating the manager instance. Generally this is host:port
+* instanceType: Type of the process. This can be one of the following types, in this case, use PARTICIPANT
+    * CONTROLLER: Process that controls the cluster, any number of controllers can be started but only one will be active at any given time
+    * PARTICIPANT: Process that performs the actual task in the distributed system
+    * SPECTATOR: Process that observes the changes in the cluster
+    * ADMIN: To carry out system admin actions
+* zkConnectString: Connection string to ZooKeeper. This is of the form host1:port1,host2:port2,host3:port3
+
+After the Helix manager instance is created, the only thing that needs to be registered is the state model factory.
+The methods of the state model will be called when controller sends transitions to the participant.  In this example, we'll use the OnlineOffline factory.  Other options include:
+
+* MasterSlaveStateModelFactory
+* LeaderStandbyStateModelFactory
+* BootstrapHandler
+
+
+```
+manager = HelixManagerFactory.getZKHelixManager(clusterName,
+                                                instanceName,
+                                                InstanceType.PARTICIPANT,
+                                                zkConnectString);
+StateMachineEngine stateMach = manager.getStateMachineEngine();
+
+//create a stateModelFactory that returns a statemodel object for each partition.
+stateModelFactory = new OnlineOfflineStateModelFactory();
+stateMach.registerStateModelFactory(stateModelType, stateModelFactory);
+manager.connect();
+```
+
+### Example State Model Factory
+
+Helix doesn\'t know what it means to change from OFFLINE\-\-\>ONLINE or ONLINE\-\-\>OFFLINE.  The following code snippet shows where you insert your system logic for these two state transitions.
+
+```
+public class OnlineOfflineStateModelFactory extends
+    StateModelFactory<StateModel> {
+  @Override
+  public StateModel createNewStateModel(String stateUnitKey) {
+    OnlineOfflineStateModel stateModel = new OnlineOfflineStateModel();
+    return stateModel;
+  }
+  @StateModelInfo(states = "{'OFFLINE','ONLINE'}", initialState = "OFFINE")
+  public static class OnlineOfflineStateModel extends StateModel {
+    @Transition(from = "OFFLINE", to = "ONLINE")
+    public void onBecomeOnlineFromOffline(Message message,
+        NotificationContext context) {
+      System.out.println("OnlineOfflineStateModel.onBecomeOnlineFromOffline()");
+
+      ////////////////////////////////////////////////////////////////////////////////////////////////
+      // Application logic to handle transition                                                     //
+      // For example, you might start a service, run initialization, etc                            //
+      ////////////////////////////////////////////////////////////////////////////////////////////////
+    }
+
+    @Transition(from = "ONLINE", to = "OFFLINE")
+    public void onBecomeOfflineFromOnline(Message message,
+        NotificationContext context) {
+      System.out.println("OnlineOfflineStateModel.onBecomeOfflineFromOnline()");
+
+      ////////////////////////////////////////////////////////////////////////////////////////////////
+      // Application logic to handle transition                                                     //
+      // For example, you might shutdown a service, log this event, or change monitoring settings   //
+      ////////////////////////////////////////////////////////////////////////////////////////////////
+    }
+  }
+}
+```

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_propstore.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_propstore.md b/website/0.6.3/src/site/markdown/tutorial_propstore.md
new file mode 100644
index 0000000..c618ea9
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_propstore.md
@@ -0,0 +1,34 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Application Property Store</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Application Property Store
+
+In this chapter, we\'ll learn how to use the application property store.
+
+### Property Store
+
+It is common that an application needs support for distributed, shared data structures.  Helix uses ZooKeeper to store the application data and hence provides notifications when the data changes.
+
+While you could use ZooKeeper directly, Helix supports caching the data with a write-through cache. This is far more efficient than reading from ZooKeeper for every access.
+
+See [HelixManager.getHelixPropertyStore](http://helix.apache.org/javadocs/0.6.3/reference/org/apache/helix/store/package-summary.html) for details.

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_rebalance.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_rebalance.md b/website/0.6.3/src/site/markdown/tutorial_rebalance.md
new file mode 100644
index 0000000..2e1a79b
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_rebalance.md
@@ -0,0 +1,181 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Rebalancing Algorithms</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Rebalancing Algorithms
+
+The placement of partitions in a distributed system is essential for the reliability and scalability of the system.  For example, when a node fails, it is important that the partitions hosted on that node are reallocated evenly among the remaining nodes. Consistent hashing is one such algorithm that can satisfy this guarantee.  Helix provides a variant of consistent hashing based on the RUSH algorithm, among others.
+
+This means given a number of partitions, replicas and number of nodes, Helix does the automatic assignment of partition to nodes such that:
+
+* Each node has the same number of partitions
+* Replicas of the same partition do not stay on the same node
+* When a node fails, the partitions will be equally distributed among the remaining nodes
+* When new nodes are added, the number of partitions moved will be minimized along with satisfying the above criteria
+
+Helix employs a rebalancing algorithm to compute the _ideal state_ of the system.  When the _current state_ differs from the _ideal state_, Helix uses it as the target state of the system and computes the appropriate transitions needed to bring it to the _ideal state_.
+
+Helix makes it easy to perform this operation, while giving you control over the algorithm.  In this section, we\'ll see how to implement the desired behavior.
+
+Helix has four options for rebalancing, in increasing order of customization by the system builder:
+
+* FULL_AUTO
+* SEMI_AUTO
+* CUSTOMIZED
+* USER_DEFINED
+
+```
+            |FULL_AUTO     |  SEMI_AUTO | CUSTOMIZED|  USER_DEFINED  |
+            ---------------------------------------------------------|
+   LOCATION | HELIX        |  APP       |  APP      |      APP       |
+            ---------------------------------------------------------|
+      STATE | HELIX        |  HELIX     |  APP      |      APP       |
+            ----------------------------------------------------------
+```
+
+
+### FULL_AUTO
+
+When the rebalance mode is set to FULL_AUTO, Helix controls both the location of the replica along with the state. This option is useful for applications where creation of a replica is not expensive.
+
+For example, consider this system that uses a MasterSlave state model, with 3 partitions and 2 replicas in the ideal state.
+
+```
+{
+  "id" : "MyResource",
+  "simpleFields" : {
+    "REBALANCE_MODE" : "FULL_AUTO",
+    "NUM_PARTITIONS" : "3",
+    "REPLICAS" : "2",
+    "STATE_MODEL_DEF_REF" : "MasterSlave",
+  }
+  "listFields" : {
+    "MyResource_0" : [],
+    "MyResource_1" : [],
+    "MyResource_2" : []
+  },
+  "mapFields" : {
+  }
+}
+```
+
+If there are 3 nodes in the cluster, then Helix will balance the masters and slaves equally.  The ideal state is therefore:
+
+```
+{
+  "id" : "MyResource",
+  "simpleFields" : {
+    "NUM_PARTITIONS" : "3",
+    "REPLICAS" : "2",
+    "STATE_MODEL_DEF_REF" : "MasterSlave",
+  },
+  "mapFields" : {
+    "MyResource_0" : {
+      "N1" : "MASTER",
+      "N2" : "SLAVE",
+    },
+    "MyResource_1" : {
+      "N2" : "MASTER",
+      "N3" : "SLAVE",
+    },
+    "MyResource_2" : {
+      "N3" : "MASTER",
+      "N1" : "SLAVE",
+    }
+  }
+}
+```
+
+Another typical example is evenly distributing a group of tasks among the currently healthy processes. For example, if there are 60 tasks and 4 nodes, Helix assigns 15 tasks to each node.
+When one node fails, Helix redistributes its 15 tasks to the remaining 3 nodes, resulting in a balanced 20 tasks per node. Similarly, if a node is added, Helix re-allocates 3 tasks from each of the 4 nodes to the 5th node, resulting in a balanced distribution of 12 tasks per node..
+
+### SEMI_AUTO
+
+When the application needs to control the placement of the replicas, use the SEMI_AUTO rebalance mode.
+
+Example: In the ideal state below, the partition \'MyResource_0\' is constrained to be placed only on node1 or node2.  The choice of _state_ is still controlled by Helix.  That means MyResource_0.MASTER could be on node1 and MyResource_0.SLAVE on node2, or vice-versa but neither would be placed on node3.
+
+```
+{
+  "id" : "MyResource",
+  "simpleFields" : {
+    "REBALANCE_MODE" : "SEMI_AUTO",
+    "NUM_PARTITIONS" : "3",
+    "REPLICAS" : "2",
+    "STATE_MODEL_DEF_REF" : "MasterSlave",
+  }
+  "listFields" : {
+    "MyResource_0" : [node1, node2],
+    "MyResource_1" : [node2, node3],
+    "MyResource_2" : [node3, node1]
+  },
+  "mapFields" : {
+  }
+}
+```
+
+The MasterSlave state model requires that a partition has exactly one MASTER at all times, and the other replicas should be SLAVEs.  In this simple example with 2 replicas per partition, there would be one MASTER and one SLAVE.  Upon failover, a SLAVE has to assume mastership, and a new SLAVE will be generated.
+
+In this mode when node1 fails, unlike in FULL_AUTO mode the partition is _not_ moved from node1 to node3. Instead, Helix will decide to change the state of MyResource_0 on node2 from SLAVE to MASTER, based on the system constraints.
+
+### CUSTOMIZED
+
+Helix offers a third mode called CUSTOMIZED, in which the application controls the placement _and_ state of each replica. The application needs to implement a callback interface that Helix invokes when the cluster state changes.
+Within this callback, the application can recompute the idealstate. Helix will then issue appropriate transitions such that _Idealstate_ and _Currentstate_ converges.
+
+Here\'s an example, again with 3 partitions, 2 replicas per partition, and the MasterSlave state model:
+
+```
+{
+  "id" : "MyResource",
+  "simpleFields" : {
+    "REBALANCE_MODE" : "CUSTOMIZED",
+    "NUM_PARTITIONS" : "3",
+    "REPLICAS" : "2",
+    "STATE_MODEL_DEF_REF" : "MasterSlave",
+  },
+  "mapFields" : {
+    "MyResource_0" : {
+      "N1" : "MASTER",
+      "N2" : "SLAVE",
+    },
+    "MyResource_1" : {
+      "N2" : "MASTER",
+      "N3" : "SLAVE",
+    },
+    "MyResource_2" : {
+      "N3" : "MASTER",
+      "N1" : "SLAVE",
+    }
+  }
+}
+```
+
+Suppose the current state of the system is 'MyResource_0' \-\> {N1:MASTER, N2:SLAVE} and the application changes the ideal state to 'MyResource_0' \-\> {N1:SLAVE,N2:MASTER}. While the application decides which node is MASTER and which is SLAVE, Helix will not blindly issue MASTER\-\-\>SLAVE to N1 and SLAVE\-\-\>MASTER to N2 in parallel, since that might result in a transient state where both N1 and N2 are masters, which violates the MasterSlave constraint that there is exactly one MASTER at a time.  Helix will first issue MASTER\-\-\>SLAVE to N1 and after it is completed, it will issue SLAVE\-\-\>MASTER to N2.
+
+### USER_DEFINED
+
+For maximum flexibility, Helix exposes an interface that can allow applications to plug in custom rebalancing logic. By providing the name of a class that implements the Rebalancer interface, Helix will automatically call the contained method whenever there is a change to the live participants in the cluster. For more, see [User-Defined Rebalancer](./tutorial_user_def_rebalancer.html).
+
+### Backwards Compatibility
+
+In previous versions, FULL_AUTO was called AUTO_REBALANCE and SEMI_AUTO was called AUTO. Furthermore, they were presented as the IDEAL_STATE_MODE. Helix supports both IDEAL_STATE_MODE and REBALANCE_MODE, but IDEAL_STATE_MODE is now deprecated and may be phased out in future versions.

http://git-wip-us.apache.org/repos/asf/helix/blob/a3a59e04/website/0.6.3/src/site/markdown/tutorial_spectator.md
----------------------------------------------------------------------
diff --git a/website/0.6.3/src/site/markdown/tutorial_spectator.md b/website/0.6.3/src/site/markdown/tutorial_spectator.md
new file mode 100644
index 0000000..e43cd6b
--- /dev/null
+++ b/website/0.6.3/src/site/markdown/tutorial_spectator.md
@@ -0,0 +1,75 @@
+<!---
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+<head>
+  <title>Tutorial - Spectator</title>
+</head>
+
+## [Helix Tutorial](./Tutorial.html): Spectator
+
+Next, we\'ll learn how to implement a __spectator__.  Typically, a spectator needs to react to changes within the distributed system.  Examples: a client that needs to know where to send a request, a topic consumer in a consumer group.  The spectator is automatically informed of changes in the _external state_ of the cluster, but it does not have to add any code to keep track of other components in the system.
+
+### Start a Connection
+
+Same as for a participant, The Helix manager is the common component that connects each system component with the cluster.
+
+It requires the following parameters:
+
+* clusterName: A logical name to represent the group of nodes
+* instanceName: A logical name of the process creating the manager instance. Generally this is host:port
+* instanceType: Type of the process. This can be one of the following types, in this case, use SPECTATOR:
+    * CONTROLLER: Process that controls the cluster, any number of controllers can be started but only one will be active at any given time
+    * PARTICIPANT: Process that performs the actual task in the distributed system
+    * SPECTATOR: Process that observes the changes in the cluster
+    * ADMIN: To carry out system admin actions
+* zkConnectString: Connection string to ZooKeeper. This is of the form host1:port1,host2:port2,host3:port3
+
+After the Helix manager instance is created, the only thing that needs to be registered is the listener.  When the ExternalView changes, the listener is notified.
+
+A spectator observes the cluster and is notified when the state of the system changes. Helix consolidates the state of entire cluster in one Znode called ExternalView.
+Helix provides a default implementation RoutingTableProvider that caches the cluster state and updates it when there is a change in the cluster.
+
+```
+manager = HelixManagerFactory.getZKHelixManager(clusterName,
+                                                instanceName,
+                                                InstanceType.SPECTATOR,
+                                                zkConnectString);
+manager.connect();
+RoutingTableProvider routingTableProvider = new RoutingTableProvider();
+manager.addExternalViewChangeListener(routingTableProvider);
+```
+
+### Spectator Code
+
+In the following code snippet, the application sends the request to a valid instance by interrogating the external view.  Suppose the desired resource for this request is in the partition myDB_1.
+
+```
+// instances = routingTableProvider.getInstances(, "PARTITION_NAME", "PARTITION_STATE");
+instances = routingTableProvider.getInstances("myDB", "myDB_1", "ONLINE");
+
+////////////////////////////////////////////////////////////////////////////////////////////////
+// Application-specific code to send a request to one of the instances                        //
+////////////////////////////////////////////////////////////////////////////////////////////////
+
+theInstance = instances.get(0);  // should choose an instance and throw an exception if none are available
+result = theInstance.sendRequest(yourApplicationRequest, responseObject);
+
+```
+
+When the external view changes, the application needs to react by sending requests to a different instance.


Mime
View raw message