0 views collected around this technical thread.
An experienced DBA explains why adding a new column in MySQL can unexpectedly raise duplicate entry errors during online DDL, detailing the role of primary and unique keys, the innodb_online_alter_log_max_size setting, and how concurrent DML operations cause the conflict.
This article details a systematic performance test of OceanBase's online DDL capabilities, measuring TPS and QPS impacts across various DDL operations, describing the test environment, scripts, results, and conclusions for reliable database schema changes.
By dropping a low‑selectivity composite index, adding targeted indexes on product_id‑sequence‑station_no and receive_time, and using online DDL with pt‑osc, the team reduced row scans and replication lag, then implemented batch deletions by primary‑key ranges, cutting delete time from minutes to seconds on a 100‑million‑row table.
While MySQL 5.7 native DDL methods (copy, inplace rebuild, index build, metadata‑only) vary in speed, space use, lock impact, and binlog output, GH‑OST consistently offers the fastest execution, lowest lock blocking, real‑time replication, albeit at roughly double storage cost, making it ideal for risk‑averse online schema changes.
The article investigates why altering a MySQL VARCHAR column from a length that fits within 255 bytes to one that exceeds this limit (e.g., VARCHAR(63) to VARCHAR(64) in utf8mb4) triggers a costly copy operation, analyzes the storage mechanics, reproduces the issue, and provides practical recommendations to avoid performance penalties.
This article explains the space requirements of MySQL Online DDL operations, analyzes temporary log, sort, and intermediate files, provides step‑by‑step test preparation, demonstrates failure and recovery scenarios, and summarizes best practices for avoiding space‑related errors.
This article analyzes why ALTER TABLE operations that extend VARCHAR columns sometimes fail in MySQL, explains the underlying Online DDL constraints, presents several real‑world problems such as default‑value handling, indexed columns, and hidden table‑rebuild quirks, and offers practical solutions for each case.
This article explains the differences between MySQL DDL and DML, describes how online DDL works—including copy and inplace algorithms, execution phases, common pitfalls, limitations, and recent MySQL 8.0 enhancements—providing practical guidance for safely altering large tables.
This article details a step‑by‑step strategy for expanding a VARCHAR column from 50 to 500 characters on a 4‑hundred‑million‑row MySQL 8.0 table, evaluating online DDL, gh‑ost, pt‑osc, and ultimately implementing a cascade replication workflow to meet a four‑hour downtime window.
This article explains how to add unique indexes to large MySQL tables using Online DDL, pt‑osc, and gh‑ost, compares their advantages and drawbacks, outlines common data‑loss risks, and provides practical hook scripts and best‑practice recommendations to mitigate those risks.
This article demonstrates how MySQL 8.0.29’s ALGORITHM=INSTANT enables instant column addition, deletion, and row‑size checks at any position in large tables, compares performance with previous versions, and discusses related XtraBackup backup considerations.
This article analyses why physical backups using Xtrabackup fail when MySQL 5.7 performs online DDL operations that skip redo logging, demonstrates the issue with real‑world logs, and presents four practical solutions—including lock‑DDL parameters, per‑table MDL locks, and disabling online DDL—to ensure consistent backups.
This article explains why deleting rows in MySQL does not shrink the underlying table file, describes the InnoDB storage mechanics that cause this behavior, and provides practical solutions such as OPTIMIZE TABLE, ALTER TABLE, and Online DDL to reclaim space and avoid table‑locking issues.
The article explains why MySQL’s DELETE command does not shrink the physical table file, describes InnoDB’s storage mechanics, and shows how to reclaim space using OPTIMIZE TABLE, ALTER TABLE, and Online DDL options.
The article explains why deleting rows in MySQL does not shrink the physical table file, describes the underlying InnoDB storage mechanisms, and provides practical methods such as OPTIMIZE TABLE, ALTER TABLE, and Online DDL to reclaim space and reorganize the table.
The article explains why deleting rows in MySQL does not shrink the .ibd file because InnoDB only marks pages as reusable, and shows how to reclaim space using OPTIMIZE TABLE, ALTER TABLE, and Online DDL options.
This article explains why deleting rows in MySQL does not decrease the physical table file size, describes the underlying InnoDB storage mechanisms, and shows how to reclaim space using OPTIMIZE TABLE, ALTER TABLE, and Online DDL options.
This article shares practical experience with TiDB Data Migration (DM), covering its background, architecture, key features, online DDL support, common error handling such as duplicate‑key issues, large‑scale import tuning, configuration limits, and cleanup recommendations for reliable MySQL‑to‑TiDB synchronization.
The article explains how MySQL's Online DDL handles VARCHAR column expansions, demonstrates performance differences across character sets using sysbench, and clarifies why certain size changes trigger in‑place alterations while others require full table copies, providing practical guidance for DBAs.
This article shares a comprehensive overview of TiDB Data Migration (DM), covering its architecture, configuration, online DDL support, common pitfalls such as duplicate‑key errors, large‑scale import tuning, version limits, and cleanup recommendations to help DBAs efficiently migrate MySQL/MariaDB workloads to TiDB.