Deploying Rackspace Cloud Big Data
The Rackspace Cloud Big Data Platform provides a scalable, robust, and complete Hadoop cluster within a few clicks. All Cloud Big Data deployments are backed by Hortonworks Data Platform (HDP). Using HDP enables Cloud Big Data to take advantage of Hadoop packages and patches provided by Hortonworks, as well as an escalation path to some of the top contributors of the core Hadoop projects.
Two deployments are available: HDP 1.3 and HDP 2.1. Both include the core Hadoop components (for example: MapReduce and HDFS) as well as Pig and Hive. The 2.1 deployment also includes Tez. Pig and Hive will take advantage Tez for data processing, by default.
Build a cluster
There are multiple methods for deploying and scaling your Hadoop cluster: the API, Cloud Control Panel, and Lava command line. API and Lava client walkthroughs are provided under Using the Python Lava Client in the Getting Started Guide.
You can create a cluster through the Cloud Control Panel as follows:
- In the Servers section of the control panel, click Big Data.
Click Create Cluster.
Fill out the following fields, and then click Create Cluster:
- Cluster Name - Specify a name to identify and distinguish the cluster
- Region - Specify the region in which to create the cluster.
- Cluster Type - Select the type of HDP deployment. For more information, see .
- Node Size - Specify the flavor of data nodes to use within the cluster.
- Number of Nodes - The number of datanodes within your cluster.
- Username - Specify a user that will be created on all nodes for access to and administration of the cluster.
- Password - Specify a password for the user.
- (Optional) SSH Key Name - If you want to access the cluster by using an SSH key, provide a name for the key.
- (Optional) SSH Key - If you chose to use an SSH key, specify the SSH public key.
- (Optional) Post-Install Script - Specify the URL of a custom script to download and run on each node after deployment of the node has completed.
- (Optional) Alternate Storage - If you want the option of accessing and storing data in Cloud Files from your cluster, select this check box.
- Username - If you are using alternate storage, specify the username of your Rackspace cloud account.
- API Key - If you are using alternate storage, specify the API key associated with your Rackspace cloud account.
- After the status changes to Active, use SSH to log in to your Gateway node’s PublicNet IP address, using the username and password that you provided in the user profile.
We recommend that you use the HDP 2.1 deployment, but if your workflow requires the older versions of Hadoop, you can use the HDP 1.3 deployment. One difference you will notice is that the HDP 2.1 deployment includes a secondary Namenode where the HDP 1.3 deployment has a stand alone Namenode.
Choosing a data node size
Cloud Big Data offers small (1.25 TB) and extra-large (10 TB) flavors. For complete specifications and pricing, see http://www.rackspace.com/cloud/big-data/pricing/
For maximum performance, choose extra-large data nodes, which take up an entire physical server to provide consistent cluster performance. If you prefer to scale your environment more granularly or have lower storage and processing needs, you can choose small data nodes.
Following are some great links for further reading about data processing as well as a data ingest method supported by Rackspace:
Rackspace Swiftfs :
Getting Data Into Your Cluster:
Apache Tez :
© 2011-2013 Rackspace US, Inc.
Except where otherwise noted, content on this site is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License
See license specifics and DISCLAIMER