ViPR 2.1 - Install a ViPR Data Services node

Table of Contents

Introduction

ViPR data services can be supported on file storage systems or on commodity nodes. When using ViPR-supported file storage systems (Isilion, VNX, or NetApp) to support ViPR data services, one or more data services nodes must be added to your ViPR controller cluster. This article describes how to install a data services node virtual appliance to a VMware vSphere environment.

If you want to install data services on commodity nodes, you should refer to the ViPR Installation and Configuration Roadmap.

Back to Top

Prerequisites

Before you can deploy a data service node you must obtain the appropriate software and you must install a license that allows you to configure and use ViPR data services.

Obtaining and installing a license for ViPR Data Services

To use ViPR Data Services you will need an Unstructured license that includes the "object" access mode. Information on obtaining and installing ViPR licenses is provide in the following article: Obtain and Upload License File to EMC ViPR.

Obtaining Data Services deployment files and HDFS support files

To install one or more data services nodes to enable ViPR Data Services, you will need to download the vipr-*-dataservice.zip from the ViPR product page on support.EMC.com.

In addition, if you intend to configure a Hadoop cluster to use ViPR HDFS storage, you will need to obtain the ViPR HDFS zip file (vipr-hdfs-<version>.zip) from support.EMC.com. This file contains a JAR file that must be deployed to each client node in the Hadoop cluster and contains tools to aid the configuration of ViPR to allow access from the Hadoop cluster.

Before you can set up a Hadoop cluster to use ViPR HDFS, you must complete the installation and configuration of Data Services. You can then refer to the instructions on configuring a Hadoop cluster to use ViPR HDFS.

Back to Top

Install Data Services node VM steps

The steps required to prepare ViPR to work with data services nodes and to deploy the data services node VMs themselves are provided in this topic.

Once the data services node install is complete, you can configure object storage as described in the following article: Configure ViPR Object Storage on ViPR-Managed File Storage Systems.

Back to Top

Configure the ViPR controller to allow access by Data Services nodes

The Controller node needs to know the data services node VMs that it should allow to connect through its firewall so that they can access its coordination services. To enable this, before you deploy data services nodes, the addresses of the data services VMs must be specified in the ViPR controller node configuration. This configuration must be in place when the data services node VM first starts. If it is not, the VMs need to be restarted, one by one, after the data node addresses are added.

Before you begin

  • You will need the System Administrator role and Security Administrator role in ViPR.

  • You must allocate IPv4 addresses for all data nodes that you intend to add. If you create a subnet for the data VMs, you can use the subnet address. Using a subnet enables you to add data VMs without having to perform this procedure each time a data services node is added.

Procedure

  1. Open a web browser window on https://ViPR_virtual_ip to run the ViPR UI.
  2. Log in to the UI as a user who has the System Administrator role and the Security Administrator role. The root user has these permissions or the Security Administrator can assign these roles to a user.
  3. Select Settings > Data Services Nodes and click on the Configuration Properties link.
    This page is designed to provide guidance when deploying and configuring a data services node VM. You can also go directly to the configuration page at Settings > Configuration.
  4. In the Network tab, Extra Node IP Addresses field of the Configuration page, enter the IPv4 address of the data services node VMs that you are planning to deploy.
    If you are adding more than one, enter all addresses in a comma separated list. Alternatively, if you have created a new subnet for the data VMs, enter the subnet address.
  5. Select Save.

Results

The ViPR controller restarts. Once restarted and the ViPR UI is restored you can go to the Administration mode and return to the Settings > Data Services Nodes page.

Back to Top

Prepare the data node VM installation files and configuration file

Before you can deploy a data services node VM, you must obtain and unzip the installation files and you must obtain the data services node configuration file.

Before you begin

  • You need access to the ViPR data node VM distribution archive file.

  • You need to be able to access the UI for the deployed ViPR virtual appliance.

  • You will need the System Administrator role and Security Administrator role in ViPR.

Procedure

  1. Identify a working location that is accessible from the vCenter where you are deploying ViPR and extract the data services node distribution archive, vipr-*-dataservice.zip, to this location. The working location does not need to be on a ViPR VM. You can use laptop storage, for example.
    The extracted files comprise a dataservice OVF and the ViPR virtual machine disk files (VMDKs): vipr-*-dataservice.ovf, vipr-*-disk1.vmdk, vipr-*-disk2.vmdk, and vipr-*-disk3.vmdk.
    To this installation file set you will add a data services node configuration file.
  2. Ensure you are logged in at the UI as a user who has the System Administrator role and the Security Administrator role. The root user has these permissions or the Security Administrator can assign these roles to a user.
  3. Select Settings > Data Services Nodes.
  4. In the Download section of the page, select Download Configuration File to download the Data Services configuration file (config.iso).
    If you have a previously downloaded .iso file, do not use it if you have made any controller node configuration changes since it was downloaded. Use a newly downloaded .iso file.

    During deployment, the ISO image is mounted by vCenter and configuration information required by the data node VM is obtained. This information comprises the addresses and ports of the controller node services that the data node needs to access.

  5. Using an OS command, file browser, or the like (not the ViPRUI), copy config.iso from the download location to the working location where the OVF file and VMDK files reside.
Back to Top

Deploy ViPR data VMs

To support data services on ViPR-managed file systems, you need to deploy one or more data VMs in addition to the ViPRController VMs. This procedure can be repeated for each data services VM.

Before you begin

  • The ViPR virtual appliance to which you are adding data services VMs must already be deployed.

  • You must have the ViPR data services node installation files and the configuration file in a working location that is accessible from the vCenter where you are deploying.

  • The IPv4 address or hostname of all data VMs must have been added to the ViPR Controller configuration using the procedure in Configure the ViPR controller to allow access by Data Services nodes.

  • In addition to having an available IPv4 address for each data nodes, you need to know the subnet mask and gateway address for the data nodes.

  • If you are deploying the data node(s) immediately after deploying the ViPR virtual appliance, wait until the ViPR virtual appliance status on the System > Dashboard page at the ViPR UI says "Stable" (this can take a few minutes).

  • Refer to the EMC ViPR Support Matrix for the CPU, memory, and storage requirements of a data VM.

Procedure

  1. Log in to vCenter using the vSphere client.
  2. From the File menu, select Deploy OVF Template....
  3. Browse to and select the dataservice OVF file, located in the working location previously created.
  4. On the OVF Template Details page, review the details about the data node VM.
  5. Accept the End User License Agreement.
  6. Enter a name for the data services node VM and select its location.
  7. Select the host or cluster that hosts the VM.
    If you selected a host or cluster before starting the deployment, you are not be offered this selection.
  8. Select a resource pool for the VM.
    If you selected a resource pool before starting the deployment, you are not be offered this selection.
    When selecting the resource pool, do not select the controller vApp.
  9. If more than one datastore is attached to the ESX host, select the datastore for your VM.
  10. Select a disk format: Thick Provision Lazy Zeroed, Thick Provision Eager Zeroed (recommended for production deployment), or Thin Provision.
  11. On the Network Mapping page, map the source network to a destination network as appropriate.
  12. Enter the following IP address properties for the VM and its network in IPv4 format.
  13. Review the selections you have made at the Ready to Complete page and select Finish.

After you finish

Once the deployment has completed successfully, start each data node VM, one at a time, and check to see that the data node VM appears in the ViPR Virtual Appliance area of the UI Dashboard: System > Dashboard

Back to Top