PGConf.Russia 2022

In June 2020, AlfaStrakhovaniye was obliged to use RSA 2.0 government services along with other insurance companies. It was the only way to sell car insurance in Russia. AlfaStrakhovaniye is a leader in car insurance, and its applications degraded under the pressure of calls. It wasn't planned. Postgres used a huge amount of CPU and memory. Nobody could tell what was going on. Degradation started and stopped for no reason. Dev Team and Postgres Pro experts solved that problem. MyBatis in Java services was the killer. Our story is about the development of applications for heavy workloads and technical ignorance while using frameworks.

Many books end with a wedding, but the reader has no idea about the future life of the heroes except that they lived happily ever after. In 2019, we successfully migrated distributed our railway application from Oracle 11g SE to vanilla PostgreSQL 11.9. But our story did not end with this successful migration - life went on, and sometimes we got startled because of "surprises". We encountered a number of problems, some of which were solved by reorganizing the data, some disappeared after we changed our stored procedures, and some got resolved after tuning the PostgreSQL parameters. Solving our problems would be impossible without the logging and profiling system built into our DB application. Our talk covers the examples of successful detection and resolving of the performance issues that occurred in our PostgreSQL-based application.

PostgreSQL has a number of peculiarities that you need to take into account not only while maintaining your database but also when designing your database schema. Experienced PostgreSQL are well aware of vacuuming process. On the web one can find tons of materials covering its internals, configurations and monitoring. Many valuable talks about vacuum were given at numerous conferences. However, we still face the common wraparound problem when the maximum possible number of transactions (xid) is reached. It happens even on databases that are relatively small in size. In my presentation, I will share a customer case that looks interesting to me. A chain of mistakes made at different stages of the database's life cycle once caused a disaster. The database fully stopped for one week, we detected a wraparound and spotted corrupted blocks. Maintenance was problematical, and we spent sleepless nights in search of a solution. We managed to achieve a local win as we finally restored the database, but it's not the end of this story, which makes it even more interesting.

Disaster-tolerant cluster implies minimal data loss in the event of the main data center disaster and switching to the backup data center. The data loss takes place due to the asynchronous replication between the main data center and the backup one. However, there's a solution that can resolve the existing problem and ensure zero data loss in the case of the main data center disaster. My talk covers this solution for zero data loss.