PGConf.Russia 2022

Tasks for migrations from the classic commercial DBMSs to open-source solutions are still relevant nowadays. Tools used to migrate Oracle applications to PostgreSQL has already proved to be efficient. But what if you have SQL Server? We want to introduce you to Babelfish, a OSS solution based on PostgreSQL 13. Babelfish supports TDS network protocol, T-SQL language, and SQL extensions specific to SQL Server. However, it isn't that simple. In our talk, we'll cover the features ensuring compatibility with SQL Server, existing problems and ways to solve them. Bonus: we'll also share the story on how we added the Babelfish support in DBeaver using the JDBC driver from Microsoft.

The plv8 extension is very underestimated, though it can help simplify development for a variety of tasks. It's been a year since we started using this extension in our projects. The main reason for its low popularity is the complexity of building. Moreover, building plv8 requires a lot of traffic and disc space for temporary files. That's why we decided to build plv8 and provide ready-made binaries freely to anyone interested, not just as an offering for our customers. First, we made the binaries available on GitHub. Then we created a Docker image for Postgres with pre-installed plv8 based on Debian. It looked too heavy to us, and we decided to check Alpine, however, building it for Alpine appeared to be more complicated than we expected. I'll explain what difficulties we faced and how we managed to overcome them. I will also list the tasks that can be solved with plv8 and provide advice on how to make the development process more convenient. I will also share the links to our Docker images for the latest Postgres 13 and Postgres 14 releases, that are compact and can be used for any types of tasks.

Previously, we have explained the internals of Avito, discussed, where and how we store your classifieds, and how they appear on the search results. In the recent 3-4 a lot has changed in Avito. We got rid of logical replication, stopped using standby servers for reads, removed nearly all stored procedures and our custom failover solution, migrated all our classifieds into a sharded DBMS, switched from the monolith app to microservices. I'll explain why we made such decisions, list some of the problems we encountered and describe the current state of our development process.

Every DBMS stores its data on a hard disc, so you may face a situation when your data on disc gets corrupted. This can happen due to a controller failure, logical or physical data corruption; there are also other reasons. If it is just the index file that gets corrupted, the index recreation command will enable you to restore data consistency in your DBMS. If a table file or a file of a system section gets corrupted, data restore is impossible. In this case, you need to invent workarounds. You can try to restore this data from backup files, copy it from the corrupted table, or find another way to solve this problem. In this talk, we'll tackle several cases of data corruption on disc and describe the options for restoring the data from the corrupted tables.