Oracle RAC 19c on VMware: Step-by-Step Guide

Hey everyone! Today, we’re diving deep into something super cool: installing Oracle RAC 19c on VMware . If you’re looking to set up a highly available and scalable database environment, you’ve come to the right place, guys. We’re going to walk through this process step-by-step, making sure you get a solid understanding of what’s going on. Oracle Real Application Clusters (RAC) is a powerful tool for ensuring your databases are always up and running, even if something goes wrong with a single server. And when you combine that with VMware’s virtualization prowess, you get a flexible and robust platform for your critical applications. This guide is designed to be comprehensive, so buckle up, and let’s get this done!

Prerequisites: Getting Your Ducks in a Row

Before we even think about starting the installation, we need to make sure we’ve got all our ducks in a row. Think of this as the foundation for our Oracle RAC 19c on VMware setup. Getting these prerequisites right is absolutely crucial for a smooth installation process. Skipping this step is like trying to build a house without a solid foundation – it’s just not going to end well, trust me. We need to ensure our VMware environment is properly configured, our virtual machines are set up correctly, and all the necessary operating system configurations are in place.

VMware Environment Setup

First off, let’s talk about your VMware environment . You’ll need a vSphere environment, obviously. Make sure your ESXi hosts are healthy and have enough resources – CPU, RAM, and importantly, storage – to handle multiple virtual machines running an Oracle database. We’re talking about at least two ESXi hosts for a proper RAC setup, to leverage VMware’s High Availability (HA) and Distributed Resource Scheduler (DRS) features. This gives us redundancy right at the hypervisor level, which is just chef’s kiss for an HA database solution. You’ll want to configure your virtual network properly. This typically involves multiple virtual networks: one for public access, one for private interconnect (the cluster communication), and maybe even one for storage. Ensure these networks are segregated and have adequate bandwidth. Also, consider using vSAN or a shared storage solution like NFS or iSCSI that is accessible by all your ESXi hosts. This shared storage is vital for Oracle RAC, as all nodes need to access the same database files.

Virtual Machine Configuration

Now, let’s get into the virtual machine configuration for our RAC nodes. You’ll need at least two VMs, and honestly, for a true RAC experience, I’d recommend three or even four. Each VM will represent a node in your Oracle RAC cluster. For each VM, you’ll need to assign a specific number of vCPUs, a good chunk of RAM (Oracle recommends a minimum, but more is always better for performance), and a decent amount of disk space. Remember, you’ll need space for the operating system, Oracle software, and the database itself. Crucially, each VM needs multiple network interfaces: one for the public network, one for the private interconnect (also known as the heartbeat or cache fusion network), and potentially others depending on your storage setup. It’s a good practice to assign static IP addresses to all these interfaces on all your VMs. For the private interconnect, ensure it’s a separate subnet from the public network and configured for maximum performance – sometimes dedicated virtual switches or even direct network passthrough can be beneficial here. Don’t forget to enable the ‘High Performance’ or ‘Balanced’ power management policy on your ESXi hosts for these VMs, as this can impact performance. Also, ensure that VMware Tools are installed and up-to-date on each VM; this is super important for proper integration and performance.

Operating System Setup

On the operating system front , we’re typically looking at a Linux environment. Oracle officially supports distributions like Oracle Linux, Red Hat Enterprise Linux, and sometimes SUSE Linux Enterprise Server. Make sure you install the latest supported version and apply all the latest patches. You’ll need to configure kernel parameters according to Oracle’s recommendations for RAC. This includes things like setting the shmmax and shmall parameters for shared memory, increasing the number of open file descriptors ( nofile ), and adjusting network-related parameters. You also need to create specific user accounts and groups, like grid for Oracle Grid Infrastructure and oracle for the Oracle database software. These users need appropriate permissions and ownership for the directories where Oracle software will be installed. SELinux should be disabled or set to permissive mode, as it can sometimes interfere with Oracle installations. Firewall rules also need to be configured to allow traffic on the necessary ports for cluster communication, listener, and database access. A shared storage setup, like ASM (Automatic Storage Management) or a clustered file system, is mandatory for RAC. You’ll need to ensure the OS can see and access this shared storage, whether it’s through iSCSI initiators, NFS mounts, or VMware’s virtual disk sharing capabilities. Finally, ensure you have root access or sudo privileges to perform all these OS-level configurations.

Installing Oracle Grid Infrastructure

Alright, now that our prerequisites are sorted, it’s time to get our hands dirty with the installation of Oracle Grid Infrastructure (GI) . This is the foundation for RAC, providing essential cluster management capabilities. Think of GI as the brain of your RAC cluster. It handles node services, interconnect communication, and manages shared storage. If GI isn’t installed and configured correctly, your RAC cluster simply won’t function. So, pay close attention here, guys!

Software Download and Preparation

First things first, you need to download the Oracle Grid Infrastructure software for your specific OS version from Oracle’s website. Make sure you download the correct version (e.g., 19c). Once downloaded, you’ll need to extract the software on the first node you plan to install. It’s good practice to create a dedicated directory for the Oracle software, often under /u01/app/oracle or similar, and ensure the grid user owns this directory and has the correct permissions. You’ll also need to create staging directories for the response files if you plan on using silent installations, which is highly recommended for consistency across nodes. Before running the installer, it’s crucial to verify that all the prerequisite OS settings we discussed earlier are indeed applied and that the necessary kernel modules are loaded. A quick check of the grid user’s environment variables, especially PATH and LD_LIBRARY_PATH , is also a good idea to ensure they point to the correct Oracle binaries and libraries. Don’t forget to check for any specific packages that Oracle’s installation checker might require – sometimes these aren’t explicitly mentioned in the main installation guides but are essential for a successful GI deployment. This preparation phase is where many installation hiccups can be avoided, so don’t rush it!

Running the Grid Infrastructure Installer

Now, let’s kick off the Grid Infrastructure installer . Log in as the grid user on your first node. Navigate to the directory where you extracted the GI software and run the runInstaller.sh script. You’ll be presented with Oracle Universal Installer (OUI). For a typical RAC setup, you’ll want to choose the ‘Clusterware installation’ option. During the installation, you’ll be prompted for several critical pieces of information. You’ll define your cluster name and node names . You’ll specify the interconnect network interfaces – this is super important for cluster communication. You’ll also configure shared storage , typically using ASM. You’ll create your ASM disk groups, specifying the disks that will be used for your database files, control files, and online redo logs. It’s common to have at least two disk groups: one for OCR (Oracle Cluster Registry) and voting disks, and another for data files. You’ll select the installation location for the Grid Infrastructure binaries (often called the Grid Home). The installer will perform various checks to ensure your system meets all the requirements. Pay close attention to any warnings or errors it flags. Once the configuration is complete, the installer will proceed with the actual installation and configuration of Clusterware and ASM. This can take a while, so grab a coffee!

Post-Installation Steps for Grid Infrastructure

After the installer finishes, there are a few essential post-installation steps . First, you need to run the root.sh script as the root user on each node. This script configures the necessary OS-level components and kernel modules. After running root.sh on all nodes, you need to start the cluster services. You can typically do this using the crsctl start cluster -all command. It’s also a good idea to verify the cluster status using crsctl stat res -t . This command shows you the status of all cluster resources, including nodes, listeners, and ASM instances. You should also check the ASM instances using asmcmd to ensure they are running and accessible. Verify that your OCR and voting disks are correctly configured and Quorum is maintained. Test the interconnect by ensuring nodes can communicate with each other reliably. If you’re using Oracle Restart for single-instance management, ensure its configuration is correct. The final verification is to ensure that all nodes are recognized as part of the cluster and that the cluster is stable. Don’t forget to check the alert logs for both Clusterware and ASM for any errors. This phase is critical to ensure your cluster foundation is solid before moving on to the database installation.

Installing Oracle Database Software

With Grid Infrastructure humming along nicely, we can now focus on the installation of the Oracle Database software itself. This is where we prepare our nodes to run the actual database instances that will form our RAC cluster.

Software Download and Preparation

Similar to Grid Infrastructure, you’ll need to download the Oracle Database software for your specific version (19c) and operating system from Oracle’s website. Extract the software on each node where you plan to install the database binaries. Again, use a consistent directory structure, typically under /u01/app/oracle (often referred to as the Oracle Home), and ensure the oracle user owns these directories with the correct permissions. It’s crucial to ensure that the oracle user has the necessary privileges and that the environment variables ( PATH , LD_LIBRARY_PATH , etc.) are correctly set for the oracle user on all nodes. A common mistake is having inconsistent Oracle Home locations or permissions across nodes, which can lead to bizarre issues later. Ensure all required packages and libraries are installed on the OS, as listed in the Oracle Database Installation Guide for your version. Perform thorough checks using Oracle’s pre-installation checker tools if available for the database software, just as you did for GI.

Running the Database Installer

Log in as the oracle user on one of the nodes. Navigate to the extracted database software directory and launch the runInstaller.sh script. You’ll be presented with OUI again. This time, select the ‘Oracle Real Application Clusters database installation’ option. You’ll need to specify the Oracle Home location (the directory where the database software will be installed). Crucially, you’ll select the RAC option and choose the nodes that will be part of this database cluster. The installer will detect the existing Grid Infrastructure and ASM configuration. You’ll need to choose the type of database installation – for a new RAC database, you’ll typically select ‘Create a new database’. You’ll then proceed through various configuration screens where you’ll define your database name (Global Database Name), specify database components to install, set passwords for administrative accounts (like SYS and SYSTEM), and configure database storage options. You’ll likely be using ASM for database files, so you’ll select the ASM disk groups you created earlier. The installer will also prompt you to configure the Oracle Net Services, including creating listeners for your RAC database, ensuring they are registered with the Clusterware. This is a pretty intensive step with lots of choices, so read carefully!

Post-Installation Steps for the Database

Once the database installer completes, you’ll need to perform some post-installation tasks . On each node, you’ll need to run the $ORACLE_HOME/root.sh script as the root user. This script links the Oracle binaries and sets up necessary OS-level configurations for the database software. After running root.sh on all nodes, you should verify that the database instances are registered with Clusterware and are starting automatically. You can check this using srvctl status database -d <your_db_name> . The srvctl command is your best friend for managing RAC databases. You should also check the listener status using lsnrctl status <listener_name> to ensure it’s running and accepting connections. Connect to the database using SQL*Plus as SYSDBA (e.g., sqlplus / as sysasm ) and verify that all instances are up and running and that you can connect to them. Check the database alert logs for any errors. It’s also a good practice to create a basic RAC configuration, such as enabling RAC-specific features or setting up instance-specific parameters. Finally, test failover by intentionally shutting down one instance or node and verifying that the database remains accessible and that connections are properly managed by Clusterware. This validation step is key to confirming your RAC database is healthy and ready for production use.

Testing and Verification

We’re almost there, folks! The final, but arguably most critical phase is testing and verification . Installing is one thing, but ensuring it works flawlessly under various conditions is another. We need to be absolutely sure our Oracle RAC 19c on VMware setup is robust and ready for prime time.

Clusterware and ASM Verification

First up, let’s confirm that Clusterware and ASM are running smoothly . Log in to each node and use crsctl stat res -t to verify that all resources (VIPs, listeners, ASM instances, database instances) are online and in a stable state. Check the Clusterware alert logs ( $GRID_HOME/log/<hostname>/alert<..>.log ) for any errors or warnings. Use asmcmd to connect to ASM and check the status of your disk groups. Ensure they are mounted and have sufficient free space. Try creating a small test table in one of your ASM disk groups to confirm data is being written correctly. Verify that the voting disk quorum is maintained. A simple crsctl query css votedisk can help here. Ensure that node applications (like the listener and ASM instance) are starting automatically upon node boot. This is a fundamental check that ensures your cluster management layer is solid.

Database Instance and Connectivity Testing

Next, we need to test the database instances and connectivity . Use srvctl status database -d <your_db_name> to ensure all database instances are running and managed by Clusterware. Connect to the database using SQL*Plus from a client machine and try connecting to each instance individually, and then connect using a RAC-aware client or connection string that utilizes the SCAN listener. The SCAN (Single Client Access Name) listener is vital for simplified client connections in RAC. Ensure the SCAN listener is registered and accessible from external clients. Test basic SQL queries, insert statements, and ensure data is written consistently across all nodes. Try connecting with different TNS aliases to verify network configurations are correct.

High Availability and Failover Testing

This is the real test of RAC – its high availability and failover capabilities . Gracefully shut down one of the RAC database instances using srvctl stop instance -d <your_db_name> -i <instance_name> . Verify that existing connections are properly managed (they might get disconnected and need to reconnect, or if configured correctly, they might failover). Check that Clusterware automatically restarts the instance on another node or brings it online once the node is back. You can also simulate a node failure by shutting down an entire VM (though do this carefully in a test environment!). Observe how Clusterware handles the failure, how resources are rebalanced, and how the database remains available. Test application failover scenarios if you have applications directly connecting and relying on specific instances. This part really shows the power of RAC and ensures your investment in HA is paying off. Remember, thorough failover testing is essential to build confidence in your RAC setup.

Conclusion

And there you have it, guys! We’ve walked through the entire process of installing Oracle RAC 19c on VMware , from the nitty-gritty prerequisites to the final HA testing. Setting up Oracle RAC on a virtualized platform like VMware might seem daunting at first, but by following these steps methodically, you can build a powerful, highly available, and scalable database solution. Remember, preparation is key , so don’t skimp on the prerequisite checks. A well-configured virtual environment and OS setup will save you countless headaches down the line. Grid Infrastructure is the backbone, so ensure it’s installed and configured flawlessly. And finally, always, always test your HA and failover scenarios thoroughly. This isn’t just about getting it installed; it’s about ensuring your data is safe and accessible when you need it most. Happy RACing!