This post was written in collaboration with Artem Danielov, CTO of FlashGrid.

We all expect our phones to make and take calls and our credit cards to process payments whether day or night. Business transactions are expected to always be completed on time, with no margin for downtime. With this, software has become the lifeblood of many, if not most, enterprises and public services.

Data collection and analysis is an integral part of ensuring quality of service for these mission-critical applications, and it is important that the availability of the system doing the collection is on par with the applications being observed. Having gaps in the captured data leads to unreliable analysis and can dramatically increase troubleshooting time. 

“We have no data from the recent outage because our event capture system was down just like the system it was supposed to monitor” is not something an enterprise DevOps lead wants to hear from a vendor.

In this blog, Lorenzo Conti, Principal Implementation Engineer, OverOps and Artem Danielov, CTO, FlashGrid, share their experience with building an event capture system in AWS that spans three availability zones.

The Task: Provide High Availability Support for OverOps Customers on Oracle RAC

OverOps’ key mission is to provide a complete look into the reliability of today’s complex IT systems. For our enterprise customers, the On-Prem deployment of the OverOps Platform is most common.

A crucial component of this solution is the persistence layer where there is a set of data that is managed in an RDBMS database. Enterprise customers have been investing in these kinds of solutions for a long time and many have opted for different standards. 

To support the wide variety of solutions in the industry, we’ve made our solution compatible with several different RDBMS servers: MySQL, PostGreSQL and, more recently, Oracle DB.

Oracle DB is well-represented in the enterprise world thanks to its long use history and high reliability. With a significant number of customers using this solution, we started working to support OracleDB with several more complex configurations, including optional RAC capabilities.

This, together with similar deployments using MySQL and PostGreSQL, makes it simpler and more manageable to deploy OverOps to continuously monitor the overall health of any enterprise application.

Expanding Scope: From Single Instance On-Prem Oracle DBs to RAC Clusters in the Cloud

During the development of OracleDB support, a key design objective for our engineering team was to enable the final OracleDB support to work not only with the traditional On-Prem OracleDB but also with some cloud and/or more complex OracleDB instances.

In this context, we explored and tested support for Amazon RDS for Oracle and other cloud variants with increasing level of complexity using optional components to make sure the implementation will satisfy an adequate range of customers.

Ultimately, we decided to explore supporting the most sophisticated flavor of Oracle DB, Real Application Cluster (RAC), set up on Amazon Web Services EC2 cloud. 

The question our team faced was whether or not it’s possible to set up an Oracle RAC database in AWS. And if it is, how practical is it?

Yes, Oracle RAC in AWS Is Easy with FlashGrid SkyCluster

During this phase of development, test and research, the team started exploring the FlashGrid SkyCluster Oracle Database HA for AWS. Ultimately, this solution allows our customers to have a cloud (AWS)-ready solution OracleDB with support for Oracle RAC at the same time. This, together with the ease of deployment of FlashGrid SkyCluster in AWS, has proven to be a very welcome solution and deployment option for our enterprise customers.

The active collaboration and joined effort between our team and the team at FlashGrid, both based in the Bay Area, was really one of the driving forces that enables OverOps Platform customers to elect to use OracleDB with optional RAC deployments OnPrem and/or in the cloud.

FlashGrid SkyCluster delivers all the high availability benefits of Oracle RAC database combined with the agility of deployment and flexibility of AWS:

  • A pre-tested solution architecture with many live enterprise mission-critical deployments
  • Easy configuration of the cluster with an online SkyCluster Launcher tool, followed by infrastructure-as-code scripted deployment in 1 hour
  • 24/7 support with cluster management and monitoring tools

The cluster was built using r4.2xlarge EC2 instances placed in different availability zones. EBS volumes attached to the instances were managed by FlashGrid Storage Fabric software to create shared storage. FlashGrid Cloud Area Network component ensured meeting network connectivity requirements for the cluster.

The Outcome: OverOps Customers Benefit from Always-On Event Registration and Analysis

OverOps, always eager to offer its customer base the best and widest selection of options and an active supporter for cloud deployment models, is really pleased to be able to offer support for Oracle RAC as its RDBMS. Try the OverOps Platform today or request a demo.

From our tests, FlashGrid SkyCluster has proven to be a convenient and cloud-ready solution that can be used as a valid alternative to On-Prem deployments.

Database High Availability in the Cloud with FlashGrid SkyCluster:
https://www.flashgrid.io/database-ha-in-cloud/

Lorenzo is a Senior Implementation Engineer at OverOps with nearly 30 years of software experience.