0 views collected around this technical thread.
MySQL employs a two‑phase commit to synchronize InnoDB’s redo log with the server’s binlog, ensuring atomicity during transaction commits and preventing data inconsistencies during crash recovery or replication, while also introducing lock contention, extra I/O, and potential blocking.
This article examines the data consistency challenges of writing to both databases and caches, outlines four common double‑write strategies, analyzes their pitfalls—especially under high concurrency—and recommends the optimal approach of writing to the database then deleting the cache, while discussing retry mechanisms, scheduled tasks, message queues, and binlog listeners.
This article explains MySQL's doublewrite buffer mechanism, how it ensures data consistency by writing to a redo log before disk, provides configuration steps to enable it, and outlines performance optimization techniques such as adjusting redo log size, separating log files, and tuning flush settings.
This article examines the challenges of data consistency verification during database migrations, reviews existing tools like pt-table-checksum, sync‑diff‑inspector and gt‑checksum, outlines essential capabilities, and proposes a detailed workflow and future enhancements for a versatile, low‑impact consistency checking solution.
The article examines how to maintain data consistency between caches and databases during dual-write operations, categorizes data into three levels, evaluates four update strategies, and proposes solutions such as delayed double deletion, message‑queue compensation, and binlog‑driven cache synchronization.
The article analyzes three data‑level classifications, examines four cache‑database write strategies, discusses their pros and cons, and presents practical solutions such as delayed double deletion, message‑queue compensation, and binlog‑driven cache updates to maintain consistency.
This article explains why cache‑database consistency is a classic challenge, compares common update‑then‑delete patterns, analyzes concurrency and master‑slave delay issues, and recommends reliable solutions such as asynchronous retries with message queues or binlog subscription to achieve eventual consistency.
This article explains the principles of MySQL semi‑synchronous replication, analyzes how data consistency is maintained during high‑availability failover, presents detailed step‑by‑step transaction flow, discusses scenarios causing GTID divergence, and offers testing methods and remediation techniques for DBA practitioners.
The article explains how micro‑service architectures create data‑consistency challenges that require distributed transaction strategies, compares strong‑consistency protocols like 2PC/3PC with eventual‑consistency approaches such as TCC and Saga patterns, and discusses their trade‑offs, implementation complexity, and suitability for real‑world scenarios like a points‑based sign‑in system.
This article explains the fundamentals of distributed transactions, the ACID properties, various consistency models (strong, weak, eventual), sharding strategies (vertical and horizontal), the CAP and BASE theories, and the practical implementations of two‑phase, three‑phase, and TCC commit protocols, highlighting their advantages and drawbacks.
The article examines tiered data storage and evaluates four cache‑synchronization strategies—updating the database before the cache, deleting the cache before updating the database, updating the cache before the database, and deleting the cache after a database update—highlighting their trade‑offs and practical solutions such as delayed double deletion, message‑queue retries, and binlog‑driven cache updates.
The article analyzes the challenges of maintaining data consistency when using both cache (local memory or Redis) and a database, classifies data by real‑time requirements, evaluates four double‑write strategies, and proposes practical solutions such as delayed double deletion, message‑queue compensation, and binlog‑driven cache updates.
This article explains the concept of anti‑entropy as a key mechanism for achieving eventual consistency in distributed database systems, illustrates how XDB Enterprise handles node failures and data drift, and shows practical examples of repairing inconsistencies using the AE service.
This article explains rigid (strong consistency) distributed transactions, detailing the XA protocol and the classic two‑phase commit (2PC) and three‑phase commit (3PC) mechanisms, their workflows, advantages, limitations, and practical considerations for database systems.
This article explains why cache‑database consistency is a classic problem, analyzes various update‑order and deletion strategies, discusses concurrency and replication delays, and recommends using the "update database then delete cache" approach combined with asynchronous retries via message queues or binlog subscription to achieve reliable eventual consistency.
The article uses a personal bank‑transfer incident to introduce local transaction concepts, explains the CAP theorem, examines the challenges of distributed transactions, and presents a message‑queue‑based design with transaction logs and compensation mechanisms to ensure consistency and idempotency.
This article explains how to guarantee data consistency when updating multiple tables or services by using local transactions for single‑node systems, two‑phase commit for distributed databases, and reliable message‑queue patterns—including RocketMQ transactional messages—to achieve eventual consistency and handle duplicate delivery.
This article analyzes how MySQL slave crashes can cause inconsistencies between master info and relay log positions during sync_master_info and sync_relay_log_info processes, presents case studies of duplicated events, explains the impact of different storage repositories (FILE vs TABLE), and recommends configuration settings to achieve server‑crash‑safe replication.
This article explains typical cache issues such as cache avalanche, cache penetration, concurrent write conflicts, and database‑cache write inconsistency, and provides practical mitigation strategies including high‑availability setups, circuit breakers, placeholder values, distributed locks, double‑delete, and cache‑aside patterns.
The article explains the drawbacks of traditional two‑phase commit distributed transactions—reduced availability, performance, and scalability—and proposes a message‑queue‑plus‑state‑table pattern with idempotent processing to achieve eventual consistency while improving system throughput.