The Ispirer team made it possible to upgrade PostgreSQL and PostGIS on short notice
Our customer is a mid-sized company that provides software for insurance firms. It offers software for policy administration, claims, billing, and analytics, enabling insurers to meet new business demands and expand their services. The company caters to the property and casualty insurance markets in the United States and Canada.
The main reason for updating versions was the impending support termination for PostgreSQL 11 in November 2023. The databases were used to work with critical products for the client's work. Although the client has a development team, their experience was not sufficient to update the version. The team tried to update the versions themselves, but encountered problems with Postgis upgrading from 2.5.5 to 3.
The project encompassed 15 databases, totaling 30 Terabytes in size. In order to update the infrastructure in time, it was necessary to select a team of professionals. Given the substantial workload, the company opted to seek out a proficient service provider experienced in handling PostGIS. As a result, they ended up approaching the Ispirer team.
The team initiated the service project. Updating PostgreSQL is usually fairly easy. However, due to the use of PostGIS, complications arose that led to errors in the database operations after the update process.
The whole project was divided into two parts: Proof of Technology (POT) and Version upgrade. In the POT phase, a concept for the update was developed that included the following:
- creating a test copy of the database
- updating computer and Postgis versions
- error detection
- error correction
- preparing a plan for production migration
- production migration
This plan was executed for each database. Throughout the upgrading process, the Ispirer team frequently contacted the client, offering suggestions for further optimization of the client's infrastructure.
During the upgrade process, Ispirer developers made a number of key changes, including:
- Switching from ip1 to gp3 disk types resulted in a more cost-effective infrastructure while increasing application speed.
- Upgrading from m5 to i6 instance types optimized infrastructure costs and improved application performance.
- Enabling automatic storage scaling minimized the risk of application downtime due to human factor.
- Modifying database parameters for autovacuum was implemented since the customer had not done this previously and it was having a negative impact on database performance.
It took 2.5 months to complete the upgrade from outdated versions of PostgreSQL and Postgis. Now the customer no longer has to worry about being left without database support at some point because of the end of life of Postgresql 11.
The client received an upgraded version of their PostgreSQL databases with the latest PostGIS release. The expertise of Ispirer professionals resulted in valuable recommendations for further optimizing the client's infrastructure. Implementing these additional measures enhanced database performance, leading to faster operations and reduced costs.