By Honglin Su-Oracle on Apr 21, 2009
We just released a new White Paper: Oracle VM High Availability – Hands-on Guide to Implementing Guest VM HA. Guest VM HA functionality provides a powerful, easy-to-manage solution for maximizing up-time for virtually any guest VM workload, without requiring any tailoring inside the VM, making it simple to set-up, use, and maintain.
This white paper focuses on best practices of the Oracle VM Guest VM High Availability (HA) design and implementation. It's complementary to the previous White Paper: Oracle VM – Creating & Maintaining a Highly Available Environments for Guest VMs, and serves as a practical guide to help customers design the HA environment and experience the benefits of Oracle VM. It provides a step-by-step guide to plan and set up the Oracle VM environment so you can implement the guest VM HA feature to assure predictable, reliable, and accurate restarting of failed VM and Servers.
To implement HA, you must create a cluster of Virtual Machine Servers in a server pool and have them managed by Oracle VM Manager or Oracle Enterprise Manager Grid Control. Some basic steps include:
1. Installing Oracle VM Server and Manager
2. Creating Shared Storage for the Server Pool
3. Enabling HA for the Server Pool
4. Adding a new Server to the Server Pool
5. Enabling HA for the Virtual Machines
The most important part is to create shared storage for the server pool. You can set up shared storage for the server pool in the following configurations:
* OCFS2 (Oracle Cluster File System) using the iSCSI (Internet SCSI) network protocol
* OCFS2 using SAN (Storage Area Network)
* NFS (Network File System)
The procedures for creating shared storage for HA are essentially the same as what's described in the Oracle VM Server User Guide for creating a shared virtual disk using the above storage configurations for live migration. But you have fewer steps to go through when creating shared storage for HA. For example, you don't need to manually modify /etc/fstab for enabling HA since the configuration files will be handled by Oracle VM server agent automatically when you run /usr/lib/ovs/ovs-makerepo utility. In addition, the startup of related cluster services (o2cb) will also be handled when you run /usr/lib/ovs/ovs-cluster-configure utility.
One of the common mistakes is that when the network is not configured properly, the cluster configuration files such as /etc/ocfs2/cluster.conf won't be propagated correctly to each server of the server pool. For example, the loopback address (127.0.0.1) may show up in the /etc/ocfs2/cluster.conf for some servers. You should verify your network settings (DNS, routing table, etc.), replace the loopback address with the public IP address for each server and make sure that the ocsf2 cluster configuration file (/etc/ocfs2/cluster.conf) be the same across all the servers within the same pool.
In summary, Oracle VM Guest VM HA functionality provides the following benefits:
* Auto-restart unexpectedly failed individual VMs on other servers in the server pool;
* Auto-restart all the guest VMs on another server in the server pool when an unexpected physical server failure occurs;
* Powerful cluster-based network- and storage heartbeat algorithms quickly and deterministically identify failed and/or isolated servers in the server pool to ensure rapid, accurate recovery;
* Sophisticated distributed lock management functionality for SAN, NFS, NAS, and iSCSI storage ensures VMs or entire servers can be rapidly restarted with no risk of data corruption.
For more information about Oracle VM and how customers are deploying it, please visit