PgConf.Russia 2016

This talk will cover "hasql", a highly-efficient library for integration of Haskell and PostgreSQL. The library provides an API for declarative programming, which is also quite flexible and terse. The talk will cover the benefits of declarative programming as well as the architectural and technical solutions behind the library, including the implementation of the PostgreSQL binary protocol. Hasql is used in PostgREST, a popular modern restful API for Postgres.

Enterprises need enterprise-level databases. The existing Postgres clustering solutions are not supported by the community. Postgres needs a community-supported cluster solution. There have been multiple attempts like Postgres-XC/XL, but they are still being developed separately and have low chance to be accepted by the community. Other solutions, like pg_shard, plproxy, FDW-based, etc. lack the notion of global transactions. We developed a Distributed Transaction Manager (DTM) as a Postgres extension to achieve global consistency over a number of Postgres instances. To demonstrate the capabilities of the DTM we present examples of distributed transaction processing using pg_shard and postgres_fdw. We hope that the proposed approach will be included into Postgres 9.6. This will make the development of the clustering solutions easier for all interested parties.

Whenever multiple users, processes, or threads are concurrently modifying data which is shared among them, problems can occur if race conditions are not handled somehow. These problems are particularly acute in a database which provide ACID semantics. A set of changes grouped into a database transaction must appear atomically, both to concurrent transactions and in terms of crash recovery. Each transaction must move the database from one consistent state (with regard to business rules) to another. For programming efficiency, each transaction must be able to be coded independently of what other transactions may happen to be running at the same time. In the event of a crash, all modifications made by transactions for which the application was notified of successful completion, and all modifications which had become visible to other transactions, must still be completed upon crash recovery. Over the years, various strategies have been employed to provide these guarantees, and sometimes the guarantees have been compromised in one way or another. This talk will cover the approaches taken to provide these guarantees or compromised variations of them, with an emphasis on the Serializable Snapshot Isolation (SSI) technique available in PostgreSQL (and so far not in any other production product). While SSI already performs faster and with higher concurrency than any other technique for managing race conditions with most common workloads, there are many opportunities for further enhancing performance, some of which would require the assistance of people expert in the various index access methods; these issues will be discussed. The talk will also present some rough ideas about how SSI techniques might be used with XTM in a distributed system.

Time will be reserved at the end of the talk for group discussion of optimizations and possible application in distributed environments.

PostgreSQL includes several index types: GiST, SP-GiST, GIN, and of course, the regular B-tree. DBAs are familiar with using each of these for specific use cases, GIN for full-text search, GiST for geometrical data, and so on, but how do they work internally? What makes them suitable for the cases they're typically used for?

In this presentation, I will walk through the internal structure of each of these index types, explaining what strengths and weaknesses each one of them have.