Real Application Clusters One Node architecture

Real Application Clusters One Node architecture

Himanshu Bansal | December 10, 2018

• database / Oracle

---

This post discusses the Oracle® Real Application Clusters (RAC) One Node feature in the Database Enterprise Edition, which was introduced with the 11g Release 2, provides enhanced high availability for single instance Oracle Databases, protecting them from both planned and unplanned downtime. The post also provides instructions for installing the Oracle Grid infrastructure, which is required to use One Node.

Introduction

The One Node feature is a single instance of Oracle RAC running on one node of the cluster while the second node is in a cold standby mode. One Node allows customers to standardize their database deployment and management, consolidate database storage, and upgrade to a full multi-node Oracle RAC database without downtime or disruption.

With RAC, you can achieve continuous availability. There is little or no latency because transparent application failover (TAF), which is a failover routing mechanism, redirects transactions to a surviving server when a server crashes or a node becomes unavailable.

This is special because TAF performs quick clone of the instance to a new node.

"If the node fails, Oracle Clusterware will automatically restart the Oracle
RAC One Node instance on another server in the cluster."

TAF automates the instance relocation without any downtime and does not require manual intervention, which is a significant benefit. It uses Omotion, which facilitates instance migration or relocation.

RAC One Node versus RAC

Traditionally, RAC is used in a multi-node architecture, with many separate instances residing on separate servers. RAC One Node is a multiple-instance of RAC running on a single node in a cluster and has a fast instance relocation feature in the event of a catastrophic or instant server failure.

RAC works as a live-live solution. The nodes in the cluster are all active and can accept connections and workloads and can work together as a single unit.

RAC One Node, as the name suggests as well, works as a live-standby solution, where only one node is active and the other nodes are available to accept the workload in case the active node goes down (planned or unplanned).

Benefits of RAC One Node

The following list shows some benefits of the RAC One Node architecture:

• Sessions can be migrated from the active node while staying online.
• The online upgrade to RAC from RAC One Node (and vice-versa) is easy.
• Exadata is supported.
• Supported by using Oracle Virtualization Manager (OVM).
• Rolling Patches to RAC provide the same interface as RAC One Node.
• A One Node database using DBCA (from 11.2.0.2) can be created easily.
• One Node is supported on the same platforms as RAC.
• The cluster failover mechanism ensures high availability.

Prerequisites for Installation (on Red Hat Enterprise Linux Server 6)

Before you install One Node, ensure that you perform the following tasks:

• As the root user, create the users and operating system (OS) groups that will own the Oracle software.

• Ensure that the date and time settings on all the nodes are set as closely as possible to the same date and time. You can use ntpd with the -x option or the Oracle Cluster Time Synchronization Service (CTSSD) to synchronize the time. Instructions on configuring ntpd with the -x option can be found in My Oracle Support ExtNote:551704.1.

• Configure Secure Shell (SSH) on each node in the cluster and ensure that you have the following rpms installed:

    binutils-2.20.51.0.2-5.11.el6.i686
    compat-libcap1-1.10-1.i686
    compat-libstdc++-33-3.2.3-69.el6.i686
    gcc-4.4.4-13.el6.i686
    gcc-c++-4.4.4-13.el6.i686
    glibc-2.12-1.7.el6.i686
    glibc-devel-2.12-1.7.el6.i686
    Ksh
    libgcc-4.4.4-13.el6.i686
    libstdc++-4.4.4-13.el6.i686
    libstdc++-devel-4.4.4-13.el6.i686
    libaio-0.3.107-10.el6.i686
    libaio-devel-0.3.107-10.el6.i686
    make-3.81-19.el6.i686
    sysstat-9.0.4-11.el6.i686
  

• Check the ASM libraries. Ensure that you have the following Oracle ASMLib downloads for Red Hat Enterprise Linux Server 6:

    oracleasm-support-2.1.8-1.el6.x86_64.rpm
    oracleasmlib-2.0.4-1.el6.x86_64.rpm
  

• Download the Grid and RDBMS installation software from the Oracle Downloads website. Choose the download binary depending on the platform on which you plan to install the software.

• Create a directory for the Oracle Grid CRS Home on both nodes. By default, Grid is installed in /u02/app.

     [grid@BOSODBD01 grid]$ cd /u02/app
     [grid@BOSODBD01 app]$ ls -lrt
  
     total 12
     drwxr-xr-x. 3 root oinstall 4096 Sep 28 00:31 grid
     drwxrwxr-x. 9 grid oinstall 4096 Sep 30 22:48 oracle
     drwxrwx---. 6 grid oinstall 4096 Oct  7 21:11 oraInventory
  
     [grid@BOSODBD01 grid]$ ls -lrt
  
     total 4
     drwxr-xr-x. 68 root oinstall 4096 Sep 28 00:36 11.2.0.4
  

Install Grid

Perform the following steps to install Grid:

STEP 1: INITIATE INSTALLATION

Open a Virtual Network computing (VNC) connection and, as a Grid user, change directories (cd) into the folder where you staged the Grid infrastructure installation zip file, unzip the file, and run ./runInstaller.

Click Skip software updates on the first installation window, then click Next. Select options as shown in the following image:

Then, click Advanced Installation in the next window.

STEP 2: ENTER NODE INFORMATION

Enter the cluster and SCAN name and ensure that they show the correct default nodes on the cluster node information screens as shown in the following images:

STEP 3: CONFIGURE AND TEST SSH CONNECTIVITY

Click SSH Connectivity to configure and test the password-less SSH connectivity by typing in the OS password for the user grid and select Setup as shown in the following images:

STEP 4: SELECT THE NETWORK

Choose the Network Interface Name and Type according to the subnet as shown in the following image. Associate one subnet with one interface only.

Select Oracle ASM as the storage option and specify directories for the data storage as shown in the following image:

STEP 5: SELECT THE GROUP

Be very careful while selecting the group, as shown in the following image:

STEP 6: SPECIFY GRID DIRECTORIES

Specify the Oracle Grid infrastructure for cluster home base,the software directory for its HOME directory, and an inventory directory as shown in the following images:

STEP 7: COMPLETE INSTALLATION

After the setup checks are finished, check Ignore All and click Next to run the installation as shown in the following image:

(Picture12.png)

After the installation completes, do not forget to run the root.sh file as shown in the following image:

STEP 8: POST-INSTALLATION CHECKS

Perform the following tasks to verify the installation:

• Run the following commands to check the Grid installation as the root user:

  ./crs_stat –t
  ./crsctl check cluster –all
  

• Run the following command to check the Grid installation as the grid user:

    crsctl query css votedisk
  

• Run the following command to create a diskgroup in $GRID_HOME/bin:

  asmca
  

Install the database

Perform the following steps to install the database:

STEP 1: INSTALL RDBMS

As the Oracle user (RDBMS software owner), start the installer by running the following commands:

# su – oracle
# cd directory-where-you-staged-the-RDBMS-installation-software
# ./runInstaller

NOTE: Ensure that the installer is run as the intended software owner user because the only supported method to change the software owner is to reinstall the database.

STEP 2: CHOOSE GRID OPTIONS

After following the database installation steps, when prompted for Grid Installation Options, select them as shown in the following image:

STEP 3: CHOOSE DATABASE CONFIGURATION OPTIONS

When prompted for the type of database configuration options and details, select them as shown in the following images and ensure that you select Oracle Automatic Storage Method, when asked for database storage:

STEP 4: SELECT DISK GROUP

The Disk Group is created after the Grid installation. Select it as shown in the following image:

Ignore any errors by checking Ignore All as shown in the following image:

RAC One Node commands

The following list shows some of the available One Node commmands:

• srvctl config database -d <database_name> (This command shows output entry as Type: RACOneNode, which signifies this is a One Node RAC)

• srvctl status database -d <database_name>

• srvctl relocate database -d <database_name> -n <hostname> -w 10 -v (This command changes the RAC One Node to the other node in the cluster. It starts the other instance of the database by shutting down the one that is currently running).

Conclusion

Oracle RAC One Node provides a significant enhancement in Oracle Database availability.

People now-a-days are looking to reduce server loads, improve instance availability, deduct workloads, decrease maintenance outages (which obstruct normal day-to-day operations), and ease database management. Oracle RAC One Node is a best-in-class solution for the databases and deployments, including those in the cloud.