Mythics Blog

Automating your Disaster Recovery Plan with Oracle Site Guard

Posted on October 25, 2017 by Erik Benner

Tags: Oracle, Oracle Database, Oracle OpenWorld, Oracle Hardware, Disaster Recovery, Oracle Site Guard, Oracle VM, Oracle WebLogic, Cloud, Cloud Computing, Oracle Cloud

Connect with Erik Benner on Twitter!

Often when planning a disaster recovery solution, the focus is on the two key metrics, Recovery Point Objective (RPO) and the Recovery Time Objective (RTO).  These metrics define how much data the business is willing to lose and how long the critical systems can be offline. Often when designing around the RTO, one critical factor which is not accounted for, is the time it takes for knowledgeable staff to become available after the disaster.

Recently, much of the Southern United States was impacted by incredibly powerful hurricanes. These storms not only disrupted data centers, but also the people and infrastructure in the impacted cities.  These circumstances make it impossible for IT staff to access the data centers either physically or through VPN, resulting in delays in restoring critical IT systems at the disaster recovery sites. This week at Oracle OpenWorld, I was talking to a senior IT Architect at an oil and gas company about his experience with how the floods in Houston cut off his access to power and cell for several days. Without his help the more junior staff that had access were significantly slower in restoring systems. There must be a better way to automate disaster recovery failover. Imagine a tool that would enable anyone with the appropriate access to implement a failover to the disaster recovery site.

Within the Oracle ecosystem there is a family of technologies called Site Guard. These technologies allow the architects to automate failover of databases, middleware and applications, making the process of recovery as simple as a few mouse clicks in Enterprise Manager. The Site Guard family of technologies do more than just automate the failover, they also enhance the monitoring ability across the entire technology stack ensuring that all the tasks required to switch between sites are performed correctly including Storage, Virtual Machine, Database, Weblogic, SOA etc. This also includes monitoring of the environment ensuring that appropriate DR action is taken as the application fails over between the primary and secondary sites.

Oracle Database Enterprise Edition and Oracle WebLogic Server architects can use the Maximum Availability Architecture (MAA) to provide high availability and disaster recovery for both the Database and WebLogic tiers. Automation of these systems require the Database Lifecycle pack, or the WebLogic Server Management pack for the middleware layer. For other use cases, like Oracle VM orchestration for Disaster Recovery, the Configuration Management Pack for Applications would need to be licensed. When licensing these options with Hard or Trusted partitions, you can use the core count of the VM layer in the stack to correctly align the license footprint to the size of the environment.

While the process of using Site Guard to initiate a failover is simple, the tool usually requires a fair amount of configuration and there is potentially an impact on the licensing. One recent change to Site Guard licensing actually makes it less expensive for the Enterprise to adopt this technology by changing how the Configuration Management Pack for Applications is licensed. This pack can control Virtual Machine disaster recovery in Oracle VM and the Oracle Private Cloud Appliance (PCA) and it is also now free for use on non-Oracle technologies. This allows the admin to enable the automatic failover of a VM from one site to another. This is useful for automating the failover of Windows applications and Linux VMs. As a note, if there are Oracle Database or Application technologies on the VM, Site Guard for database or middleware will still need to be licensed. Site Guard technology for Oracle VM is easiest to implement when paired with an Oracle ZFS appliance, as the tool comes pre integrated with the storage array. However, other storage arrays can be leveraged with some advanced scripting.

For more information or assistance with planning and automating your Disaster recovery strategy, please contact us or reach out to your Mythics sales rep.

Erik Benner, Vice President, Enterprise Transformation, Mythics Inc.

Comments

  • ! No comments yet

Leave a Comment