Уровень 0 · материалов: 7
В кластер включаются документы о переходе с одной системы управления базами данных на другую или отказе от одной в пользу другой, но не включаются документы о тестировании производительности конкретной распределенной системы.
Общие признаки: переход на PostgreSQL, замена других СУБД на PostgreSQL, миграция с PostgreSQL на MySQL, сравнение реляционных и NoSQL баз данных, масштабируемость и технический долг баз данных
Группа выше: PostgreSQL
Смысл: The main idea is that for many growing applications, the supposed flexibility of a schemaless NoSQL database like MongoDB can become a liability, and migrating to a robust relational database like PostgreSQL can significantly improve system stability, data integrity, and performance.
Olery migrated its data infrastructure from MongoDB to PostgreSQL to solve stability issues and schema inconsistencies, resulting in significantly faster API response times and better data reliability.
Смысл: The main idea is that the perceived flexibility of NoSQL databases like MongoDB can be a trap for developers who lack discipline or a clear data model, eventually leading to significant technical debt that necessitates a migration to a relational database like PostgreSQL.
The author details how relying on MongoDB's schema-less nature without strict DTOs and proper architectural planning led to a maintenance nightmare, eventually forcing a migration to PostgreSQL.
Смысл: The text explains the technical motivations behind Uber's decision to migrate its core data storage from PostgreSQL to MySQL. It highlights that while Postgres is powerful, its architecture regarding immutable rows and physical replication created scaling bottlenecks in write performance, network bandwidth, and operational maintenance for Uber's specific high-growth needs.
Uber migrated from PostgreSQL to MySQL to solve critical issues with write amplification, replication bandwidth, and operational scaling bottlenecks.
Смысл: The text explains the technical failures Uber experienced with PostgreSQL in high-load OLTP environments, contrasting its architecture with MySQL's InnoDB to demonstrate why MySQL was more suitable for their scale. The main idea is that architectural choices regarding data storage, MVCC, and replication have profound impacts on production stability and performance.
A technical analysis of why Uber migrated from PostgreSQL back to MySQL, focusing on architectural flaws in PostgreSQL's MVCC, replication, and caching under heavy OLTP loads.
Смысл: The main idea is to document the technical and organizational journey of migrating a massive, high-load metadata store from a proprietary system (Oracle) to an open-source one (PostgreSQL) to reduce costs and technical debt.
Yandex.Mail successfully migrated 300TB of metadata from Oracle to PostgreSQL over three years to cut licensing costs and refactor legacy code, achieving better scalability and deployment speed.
Смысл: The text describes how Lingualeo overcame the stagnation of a 10-year-old legacy monolith by rebuilding its backend from scratch using PostgreSQL. The main idea is that shifting business logic to the database and flattening management structures can lead to massive gains in scalability, security, and development speed.
Lingualeo scaled its service to handle 100k concurrent users by replacing a legacy PHP monolith with a PostgreSQL-centric architecture and flattening its management structure.
Смысл: GitLab is dropping MySQL support to reduce technical debt, accelerate development speed, and optimize infrastructure costs by standardizing on PostgreSQL, which is already used by the vast majority of its user base.
GitLab is ending MySQL support in version 12.1 due to technical limitations and high maintenance costs, urging all users to migrate to PostgreSQL.