PgConf.Russia 2020

If you care about Postgres performance, there are a number of hardware acceleration options to help with different use cases. Intel Optane DC persistent memory creates new tier in data hierarchy allowing developers to utilize performance of traditional memory combining with volume and persistency of block storage devices. Unlike traditional DRAM-only in-memory systems, where memory is small, expensive, and volatile, Intel Optane DC persistent memory makes it possible to run larger Postgres databases (terabytes) in memory for higher performance. FPGAs are integrated circuits that can be reprogrammed dynamically to accelerate a specific workload such as SQL execution and data compression. FPGA accelerators extend Postgres with hundreds of SQL reader and writer processes that work in parallel on the FPGA. It’s similar to adding hundreds new cores to boost parallel processing on your server.

From the very beginning, PostgreSQL in Avito has been solving very important tasks. All the main architecture components are built around DBMS. For over 10 years, the project has been actively evolving, and the infrastructure and architecture have changed a lot.

The talk will start with an overview of how PostgreSQL infrastructure and architecture have advanced in Avito over time and which challenges have been successfully resolved.

Then we'll discuss PostgreSQL usage scenarios in Avito in 2020: microservice architecture, sharding, hosting multiple databases on a single server instance, DBaaS (Database discovery, access control, failover, backup, archive, resource sharing, etc.), integration, and team evolution.
And finally, I'll share our backlog and wishlist.

To build a decent query plan, the optimizer has to understand statistical characteristics of underlying data. It is interesting to observe how the structure of the collected information became more complicated over time: what the optimizer relied on back in its early days and what is at his disposal now with the release of the 12th version. We will also talk about how and when statistics are collected, how to manage this process and whether it is necessary to think about it at all.

200 bases, several clusters, several terabytes of data Share our experience setting up and using patroni cluster DBMS Cluster on Linux, 1C server for windows. We use: PostgreSQL assembly for 1C, Patroni, Consul, Consul dns, Commvault, Ansible Vagrant file and Ansible playbook with roles attached.