It's common for enterprise companies to make the transition from one technology to another—and that journey depends on several contributing factors. This blog post highlights how EDB customer BG Phoenics did just that and migrated their database infrastructure from DB2 to EDB. With the traditional DB2 RDBMS setup, their initial hurdles were tied to complex management of a growing database, complicated licensing processes, lengthy configurations, and system deployment setup issues. Therefore, the goal was to achieve the following outcomes after this migration from DB2 to EDB.
- Deployment of overall database infrastructure with cluster HA setup, scalable architecture, and containerization.
- Overall high throughout and stable infrastructure, quick backups to S3 storage, performance metrics, and measurement.
- Powerful yet easy to access a content management system with around 1.5 billion documents.
- Configuring better alternatives for a mission-critical solution suite and monitoring system.
- Experience better licensing terms and dependencies for overall technical product stack.
The best methods for the database are always contingent on multiple factors—foremost being the performance and security of the overall database structure. Provided this, EDB proved to be a viable database system to fulfill all kinds of enterprise-level database needs. EDB provides compatibility with other databases, offers simplified licensing terms, and supports customer support programs, performance optimization plans, and other project planning and tracking features.
Although migration from DB2 to EDB is considered an easy process, it does not come without its own challenges. There is no direct methodology to perform this migration. A manual converter can be written to perform this migration including data migration with all the constraints such as typing and sizes. Later, this overall process can be automated to attempt data transformation, moreover, to expeditiously run the migration, batches were run in parallel.
Technology and Requirements
EDB private cloud was used to automate everything and then deployment was made to puppet using the BASH shell scripts. There were dedicated VMs for each DB and HA set up for the overall operations.
Problems and Solutions
There were various problems to solve throughout the migration journey, including: user roles, security constraints, proper documentation, setting up low-level environment variables, Postgres file configuration, and network and EFM setting. The comparison was performed after the migration and significant improvement was observed. There was a noticeable difference in overall performance—around 10-20% compared to the previous infrastructure of DB2. Deployment was faster, setup for HA setup was improved, backup and monitoring processes were smooth and optimized.
Several parameters are involved when it comes to evaluating a transition to its success, initial wins, and fulfillment of the listed challenges are the foremost to consider. There was an easy to follow licensing model, scalable, and stable architecture for deployment and performance testing. The infrastructure was compatible with large-size databases and their growing needs. Moreover, end to end automated provisioning got in place for a single HA setup and private cloud management.
In addition, seamless handling of Day 2 operations such as recalculation of resources, performance evaluation, and resilient online deployments with no downtime was possible.
There were some challenges throughout the migration process with lessons for the team to learn, yet it was a valuable and productive transition. This migration fulfilled the practical purpose of introducing ease in automating the end-to-end provisioning and fast deployment workflow. Mitigating the chances of reworking and human intervention in low-level Day2 operations. Performance calculation and metric tracking were achieved. And scalable expansion to cloud service providers was planned after overseeing the future requirements.
To learn more about migrating to Postgres and EDB, you can watch the session from Postgres Build 2020 on demand!