


Implementing MySQL Online Schema Changes with gh-ost or pt-online-schema-change
Aug 02, 2025 am 12:25 AMHow to choose gh-ost or pt-online-schema-change? 1. pt-online-schema-change belongs to Percona Toolkit, has a long history and good community support; 2. gh-ost is lighter and supports triggerless mode, suitable for high concurrency or large table scenarios. Core process during use: 1. Create a new table and apply a new schema; 2. Copy the original table data; 3. Synchronize incremental changes (trigger or binlog); 4. Replace the original table. Notes include: 1. Ensure the index and foreign keys are correct; 2. Pay attention to the short locks in the switching phase; 3. Leave enough disk space; 4. Monitor copy delays. Common error checks: 1. Check lock waiting and deadlock; 2. Check for unique key conflicts; 3. Confirm the account permissions; 4. gh-ost requires confirmation that binlog is ROW mode. In addition, peak business operations should be avoided, and impact assessments should be made, environmental drills should be tested, and rollback plans should be prepared.
Modifying the MySQL table structure online is a common requirement, especially in production environments. Using ALTER TABLE
directly can easily lock tables, affecting service availability. At this time, you need to use tools like gh-ost or pt-online-schema-change to achieve "online" changes.

Both tools can complete DDL operations without locking tables, but the usage methods and principles are slightly different. Let’s talk about their practical application and precautions from several perspectives.
How to choose: gh-ost or pt-online-schema-change?
Both are based on the idea of "creating temporary table data migration trigger synchronization", but they are different in detail processing.

- pt-online-schema-change is part of Percona Toolkit. It has a longer history, good community support, and relatively comprehensive functions.
- gh-ost is an open source tool for GitHub. It is lighter in design and supports triggerless mode (implemented through binlog), which has more advantages for certain high concurrency or large table scenarios.
If your environment allows triggers, you can choose both; if you want to reduce the impact on the main library, or if you cannot create triggers, gh-ost is more suitable.
Key steps and precautions when using
No matter which tool it is, the core process is:

- Create a new table and apply a new schema change
- Copy the original table data to the new table
- Synchronize incremental changes to the original table (via trigger or binlog)
- Replace the original table
But in actual implementation, you should pay attention to these points:
- Index and Foreign Keys : Make sure the new table structure is correct, especially if the foreign key constraints should be preserved or adjusted.
- Row lock problem : Although the table will not be locked for a long time, there will be short lock writes during the switching phase.
- Disk space : It will take up extra space during operation, especially large tables, remember to reserve them.
- Replication Delay : If replication is enabled, binlog synchronization may cause delays and requires monitoring.
Common errors and troubleshooting suggestions
When encountering problems during use, you can start from these aspects:
- Check whether there are locks waiting or deadlocks occur (via
SHOW ENGINE INNODB STATUS
) - Check if synchronization fails due to unique key conflicts
- Ensure that the database account has sufficient permissions (such as creating tables, deleting tables, triggers, etc.)
- If you are using gh-ost, make sure that the binlog format is ROW mode and there are correct filtering rules.
Also, do not operate during peak business periods to avoid affecting performance.
Beyond the tool: What else do you need to consider?
In addition to technical means, there are a few things you should do before going online:
- Evaluate the scope of changes in advance, such as whether it involves index reconstruction and whether it will affect the query plan
- Perform the test environment to observe resource consumption and time consumption
- There is a rollback solution, and it can be quickly recovered if there is an error.
Some companies will encapsulate a layer of automated scripts, combining health checks and notification mechanisms to simplify such operations.
Basically that's it. Both tools can solve the pain points of online DDL well. The key is to understand their working mechanism and use it in the right scenario.
The above is the detailed content of Implementing MySQL Online Schema Changes with gh-ost or pt-online-schema-change. For more information, please follow other related articles on the PHP Chinese website!

Hot AI Tools

Undress AI Tool
Undress images for free

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics

TosecurelyconnecttoaremoteMySQLserver,useSSHtunneling,configureMySQLforremoteaccess,setfirewallrules,andconsiderSSLencryption.First,establishanSSHtunnelwithssh-L3307:localhost:3306user@remote-server-Nandconnectviamysql-h127.0.0.1-P3307.Second,editMyS

ForeignkeysinMySQLensuredataintegritybyenforcingrelationshipsbetweentables.Theypreventorphanedrecords,restrictinvaliddataentry,andcancascadechangesautomatically.BothtablesmustusetheInnoDBstorageengine,andforeignkeycolumnsmustmatchthedatatypeoftherefe

mysqldump is a common tool for performing logical backups of MySQL databases. It generates SQL files containing CREATE and INSERT statements to rebuild the database. 1. It does not back up the original file, but converts the database structure and content into portable SQL commands; 2. It is suitable for small databases or selective recovery, and is not suitable for fast recovery of TB-level data; 3. Common options include --single-transaction, --databases, --all-databases, --routines, etc.; 4. Use mysql command to import during recovery, and can turn off foreign key checks to improve speed; 5. It is recommended to test backup regularly, use compression, and automatic adjustment.

Turn on MySQL slow query logs and analyze locationable performance issues. 1. Edit the configuration file or dynamically set slow_query_log and long_query_time; 2. The log contains key fields such as Query_time, Lock_time, Rows_examined to assist in judging efficiency bottlenecks; 3. Use mysqldumpslow or pt-query-digest tools to efficiently analyze logs; 4. Optimization suggestions include adding indexes, avoiding SELECT*, splitting complex queries, etc. For example, adding an index to user_id can significantly reduce the number of scanned rows and improve query efficiency.

When handling NULL values ??in MySQL, please note: 1. When designing the table, the key fields are set to NOTNULL, and optional fields are allowed NULL; 2. ISNULL or ISNOTNULL must be used with = or !=; 3. IFNULL or COALESCE functions can be used to replace the display default values; 4. Be cautious when using NULL values ??directly when inserting or updating, and pay attention to the data source and ORM framework processing methods. NULL represents an unknown value and does not equal any value, including itself. Therefore, be careful when querying, counting, and connecting tables to avoid missing data or logical errors. Rational use of functions and constraints can effectively reduce interference caused by NULL.

To reset the root password of MySQL, please follow the following steps: 1. Stop the MySQL server, use sudosystemctlstopmysql or sudosystemctlstopmysqld; 2. Start MySQL in --skip-grant-tables mode, execute sudomysqld-skip-grant-tables&; 3. Log in to MySQL and execute the corresponding SQL command to modify the password according to the version, such as FLUSHPRIVILEGES;ALTERUSER'root'@'localhost'IDENTIFIEDBY'your_new

To view the size of the MySQL database and table, you can query the information_schema directly or use the command line tool. 1. Check the entire database size: Execute the SQL statement SELECTtable_schemaAS'Database',SUM(data_length index_length)/1024/1024AS'Size(MB)'FROMinformation_schema.tablesGROUPBYtable_schema; you can get the total size of all databases, or add WHERE conditions to limit the specific database; 2. Check the single table size: use SELECTta

Character set and sorting rules issues are common when cross-platform migration or multi-person development, resulting in garbled code or inconsistent query. There are three core solutions: First, check and unify the character set of database, table, and fields to utf8mb4, view through SHOWCREATEDATABASE/TABLE, and modify it with ALTER statement; second, specify the utf8mb4 character set when the client connects, and set it in connection parameters or execute SETNAMES; third, select the sorting rules reasonably, and recommend using utf8mb4_unicode_ci to ensure the accuracy of comparison and sorting, and specify or modify it through ALTER when building the library and table.
