<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"><title>Appendix A. MySQL 5.0 Frequently Asked Questions</title><link rel="stylesheet" href="mysql-html.css" type="text/css"><meta name="generator" content="DocBook XSL Stylesheets V1.69.1"><link rel="start" href="index.html" title="MySQL 5.0 Reference Manual"><link rel="up" href="index.html" title="MySQL 5.0 Reference Manual"><link rel="prev" href="extending-mysql.html" title="Chapter 26. Extending MySQL"><link rel="next" href="error-handling.html" title="Appendix B. Errors, Error Codes, and Common Problems"></head><body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="3" align="center">Appendix A. MySQL 5.0 Frequently Asked Questions</th></tr><tr><td width="20%" align="left"><a accesskey="p" href="extending-mysql.html">Prev</a> </td><th width="60%" align="center"> </th><td width="20%" align="right"> <a accesskey="n" href="error-handling.html">Next</a></td></tr></table><hr></div><div class="appendix" lang="en"><div class="titlepage"><div><div><h2 class="title"><a name="faqs"></a>Appendix A. MySQL 5.0 Frequently Asked Questions</h2></div></div></div><div class="toc"><p><b>Table of Contents</b></p><dl><dt><span class="section"><a href="faqs.html#faqs-general">A.1. MySQL 5.0 FAQ — General</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-storage-engines">A.2. MySQL 5.0 FAQ — Storage Engines</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-sql-modes">A.3. MySQL 5.0 FAQ — Server SQL Mode</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-stored-procs">A.4. MySQL 5.0 FAQ — Stored Procedures</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-triggers">A.5. MySQL 5.0 FAQ — Triggers</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-views">A.6. MySQL 5.0 FAQ — Views</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-information-schema">A.7. MySQL 5.0 FAQ — <code class="literal">INFORMATION_SCHEMA</code></a></span></dt><dt><span class="section"><a href="faqs.html#faqs-migration">A.8. MySQL 5.0 FAQ — Migration</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-security">A.9. MySQL 5.0 FAQ — Security</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-mysql-cluster">A.10. MySQL 5.0 FAQ — MySQL Cluster</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-cjk">A.11. MySQL 5.0 FAQ — MySQL Chinese, Japanese, and Korean Character Sets</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-connectors-apis">A.12. MySQL 5.0 FAQ — Connectors & APIs</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-replication">A.13. MySQL 5.0 FAQ — Replication</a></span></dt><dt><span class="section"><a href="faqs.html#faqs-mysql-drbd-heartbeat">A.14. MySQL 5.0 FAQ — MySQL, DRBD, and Heartbeat</a></span></dt><dd><dl><dt><span class="section"><a href="faqs.html#faqs-drbd">A.14.1. Distributed Replicated Block Device</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-linux-heartbeat">A.14.2. Linux Heartbeat</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-architecture">A.14.3. DRBD Architecture</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-mysql-replication-scale">A.14.4. DRBD and MySQL Replication</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-file-systems">A.14.5. DRBD and File Systems</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-lvm">A.14.6. DRBD and LVM</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-virtualization">A.14.7. DRBD and Virtualization</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-security">A.14.8. DRBD and Security</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-system-requirements">A.14.9. DRBD and System Requirements</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-support-consulting">A.14.10. DBRD and Support and Consulting</a></span></dt></dl></dd></dl></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-general"></a>A.1. MySQL 5.0 FAQ — General</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-1-1">28.1.1: </a> When did MySQL 5.0 become production-ready (GA)? </p></li><li><p><a href="faqs.html#qandaitem-28-1-2">28.1.2: </a> Can MySQL 5.0 do subqueries? </p></li><li><p><a href="faqs.html#qandaitem-28-1-3">28.1.3: </a> Can MySQL 5.0 peform multiple-table inserts, updates, and deletes? </p></li><li><p><a href="faqs.html#qandaitem-28-1-4">28.1.4: </a> Does MySQL 5.0 have a Query Cache? Does it work on Server, Instance or Database? </p></li><li><p><a href="faqs.html#qandaitem-28-1-5">28.1.5: </a> Does MySQL 5.0 have Sequences? </p></li><li><p><a href="faqs.html#qandaitem-28-1-6">28.1.6: </a> Does MySQL 5.0 have a <a href="functions.html#function_now"><code class="literal">NOW()</code></a> function with fractions of seconds? </p></li><li><p><a href="faqs.html#qandaitem-28-1-7">28.1.7: </a> Does MySQL 5.0 work with multi-core processors? </p></li><li><p><a href="faqs.html#qandaitem-28-1-8">28.1.8: </a> Is there a hot backup tool for MyISAM like InnoDB Hot Backup? </p></li><li><p><a href="faqs.html#qandaitem-28-1-9">28.1.9: </a> Have there been there any improvements in error reporting when foreign keys fail? Does MySQL now report which column and reference failed? </p></li><li><p><a href="faqs.html#qandaitem-28-1-10">28.1.10: </a> Can MySQL 5.0 perform ACID transactions? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-1-1"></a><span class="bold"><strong>28.1.1: </strong></span><span class="bold"><strong> When did MySQL 5.0 become production-ready (GA)? </strong></span></p><p> MySQL 5.0.15 was released for production use on 19 October 2005. We are now working on MySQL 5.1, which is currently in beta. </p><p><a name="qandaitem-28-1-2"></a><span class="bold"><strong>28.1.2: </strong></span><span class="bold"><strong> Can MySQL 5.0 do subqueries? </strong></span></p><p> Yes. See <a href="sql-syntax.html#subqueries" title="12.2.8. Subquery Syntax">Section 12.2.8, “Subquery Syntax”</a>. </p><p><a name="qandaitem-28-1-3"></a><span class="bold"><strong>28.1.3: </strong></span><span class="bold"><strong> Can MySQL 5.0 peform multiple-table inserts, updates, and deletes? </strong></span></p><p> Yes. For the syntax required to perform multiple-table updates, see <a href="sql-syntax.html#update" title="12.2.10. UPDATE Syntax">Section 12.2.10, “<code class="literal">UPDATE</code> Syntax”</a>; for that required to perform multiple-table deletes, see <a href="sql-syntax.html#delete" title="12.2.1. DELETE Syntax">Section 12.2.1, “<code class="literal">DELETE</code> Syntax”</a>. </p><p> A multiple-table insert can be accomplished using a trigger whose <code class="literal">FOR EACH ROW</code> clause contains multiple <code class="literal">INSERT</code> statements within a <code class="literal">BEGIN ... END</code> block. See <a href="triggers.html#using-triggers" title="19.3. Using Triggers">Section 19.3, “Using Triggers”</a>. </p><p><a name="qandaitem-28-1-4"></a><span class="bold"><strong>28.1.4: </strong></span><span class="bold"><strong> Does MySQL 5.0 have a Query Cache? Does it work on Server, Instance or Database? </strong></span></p><p> Yes. The query cache operates on the server level, caching complete result sets matched with the original query string. If an exactly identical query is made (which often happens, particularly in web applications), no parsing or execution is necessary; the result is sent directly from the cache. Various tuning options are available. See <a href="optimization.html#query-cache" title="6.5.4. The MySQL Query Cache">Section 6.5.4, “The MySQL Query Cache”</a>. </p><p><a name="qandaitem-28-1-5"></a><span class="bold"><strong>28.1.5: </strong></span><span class="bold"><strong> Does MySQL 5.0 have Sequences? </strong></span></p><p> No. However, MySQL has an <code class="literal">AUTO_INCREMENT</code> system, which in MySQL 5.0 can also handle inserts in a multi-master replication setup. With the <code class="option">--auto-increment-increment</code> and <code class="option">--auto-increment-offset</code> startup options, you can set each server to generate auto-increment values that don't conflict with other servers. The <code class="option">--auto-increment-increment</code> value should be greater than the number of servers, and each server should have a unique offset. </p><p><a name="qandaitem-28-1-6"></a><span class="bold"><strong>28.1.6: </strong></span><span class="bold"><strong> Does MySQL 5.0 have a <a href="functions.html#function_now"><code class="literal">NOW()</code></a> function with fractions of seconds? </strong></span></p><p> No. This is on the MySQL roadmap as a “<span class="quote">rolling feature</span>”. This means that it is not a flagship feature, but will be implemented, development time permitting. Specific customer demand may change this scheduling. </p><p> However, MySQL does parse time strings with a fractional component. See <a href="data-types.html#time" title="10.3.2. The TIME Type">Section 10.3.2, “The <code class="literal">TIME</code> Type”</a>. </p><p><a name="qandaitem-28-1-7"></a><span class="bold"><strong>28.1.7: </strong></span><span class="bold"><strong> Does MySQL 5.0 work with multi-core processors? </strong></span></p><p> Yes. MySQL is fully multi-threaded, and will make use of multiple CPUs, provided that the operating system supports them. </p><p><a name="qandaitem-28-1-8"></a><span class="bold"><strong>28.1.8: </strong></span><span class="bold"><strong> Is there a hot backup tool for MyISAM like InnoDB Hot Backup? </strong></span></p><p> This is currently under development for a future MySQL release. </p><p><a name="qandaitem-28-1-9"></a><span class="bold"><strong>28.1.9: </strong></span><span class="bold"><strong> Have there been there any improvements in error reporting when foreign keys fail? Does MySQL now report which column and reference failed? </strong></span></p><p> The foreign key support in <code class="literal">InnoDB</code> has seen improvements in each major version of MySQL. Foreign key support generic to all storage engines is scheduled for MySQL 6.x; this should resolve any inadequacies in the current storage engine specific implementation. </p><p><a name="qandaitem-28-1-10"></a><span class="bold"><strong>28.1.10: </strong></span><span class="bold"><strong> Can MySQL 5.0 perform ACID transactions? </strong></span></p><p> Yes. All current MySQL versions support transactions. The <code class="literal">InnoDB</code> storage engine offers full ACID transactions with row-level locking, multi-versioning, non-locking repeatable reads, and all four SQL standard isolation levels. </p><p> The <code class="literal">NDB</code> storage engine supports the <code class="literal">READ COMMITTED</code> transaction isolation level only. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-storage-engines"></a>A.2. MySQL 5.0 FAQ — Storage Engines</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-2-1">28.2.1: </a> Where can I obtain complete documentation for MySQL storage engines? </p></li><li><p><a href="faqs.html#qandaitem-28-2-2">28.2.2: </a> Are there any new storage engines in MySQL 5.0? </p></li><li><p><a href="faqs.html#qandaitem-28-2-3">28.2.3: </a> Have any storage engines been removed in MySQL 5.0? </p></li><li><p><a href="faqs.html#qandaitem-28-2-4">28.2.4: </a> What are the unique benefits of the <code class="literal">ARCHIVE</code> storage engine? </p></li><li><p><a href="faqs.html#qandaitem-28-2-5">28.2.5: </a> Do the new features in MySQL 5.0 apply to all storage engines? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-2-1"></a><span class="bold"><strong>28.2.1: </strong></span><span class="bold"><strong> Where can I obtain complete documentation for MySQL storage engines? </strong></span></p><p> See <a href="storage-engines.html" title="Chapter 13. Storage Engines">Chapter 13, <i>Storage Engines</i></a>. That chapter contains information about all MySQL storage engines except for the <code class="literal">NDB</code> storage engine used for MySQL Cluster; <code class="literal">NDB</code> is covered in <a href="mysql-cluster.html" title="Chapter 16. MySQL Cluster">Chapter 16, <i>MySQL Cluster</i></a>. </p><p><a name="qandaitem-28-2-2"></a><span class="bold"><strong>28.2.2: </strong></span><span class="bold"><strong> Are there any new storage engines in MySQL 5.0? </strong></span></p><p> Yes. The <code class="literal">FEDERATED</code> storage engine, new in MySQL 5.0, allows the server to access tables on other (remote) servers. See <a href="storage-engines.html#federated-storage-engine" title="13.7. The FEDERATED Storage Engine">Section 13.7, “The <code class="literal">FEDERATED</code> Storage Engine”</a>. </p><p><a name="qandaitem-28-2-3"></a><span class="bold"><strong>28.2.3: </strong></span><span class="bold"><strong> Have any storage engines been removed in MySQL 5.0? </strong></span></p><p> Yes. MySQL 5.0 no longer supports the <code class="literal">ISAM</code> storage engine. If you have any existing <code class="literal">ISAM</code> tables from previous versions of MySQL, you should convert these to <code class="literal">MyISAM</code> before upgrading to MySQL 5.0. </p><p><a name="qandaitem-28-2-4"></a><span class="bold"><strong>28.2.4: </strong></span><span class="bold"><strong> What are the unique benefits of the <code class="literal">ARCHIVE</code> storage engine? </strong></span></p><p> The <code class="literal">ARCHIVE</code> storage engine is ideally suited for storing large amounts of data without indexes; it has a very small footprint, and performs selects using table scans. See <a href="storage-engines.html#archive-storage-engine" title="13.8. The ARCHIVE Storage Engine">Section 13.8, “The <code class="literal">ARCHIVE</code> Storage Engine”</a>, for details. </p><p><a name="qandaitem-28-2-5"></a><span class="bold"><strong>28.2.5: </strong></span><span class="bold"><strong> Do the new features in MySQL 5.0 apply to all storage engines? </strong></span></p><p> The general new features such as views, stored procedures, triggers, <code class="literal">INFORMATION_SCHEMA</code>, precision math (<code class="literal">DECIMAL</code> column type), and the <code class="literal">BIT</code> column type, apply to all storage engines. There are also additions and changes for specific storage engines. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-sql-modes"></a>A.3. MySQL 5.0 FAQ — Server SQL Mode</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-3-1">28.3.1: </a> What are server SQL modes? </p></li><li><p><a href="faqs.html#qandaitem-28-3-2">28.3.2: </a> How many server SQL modes are there? </p></li><li><p><a href="faqs.html#qandaitem-28-3-3">28.3.3: </a> How do you determine the server SQL mode? </p></li><li><p><a href="faqs.html#qandaitem-28-3-4">28.3.4: </a> Is the mode dependent on the database or connection? </p></li><li><p><a href="faqs.html#qandaitem-28-3-5">28.3.5: </a> Can the rules for strict mode be extended? </p></li><li><p><a href="faqs.html#qandaitem-28-3-6">28.3.6: </a> Does strict mode impact performance? </p></li><li><p><a href="faqs.html#qandaitem-28-3-7">28.3.7: </a> What is the default server SQL mode when My SQL 5.0 is installed? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-3-1"></a><span class="bold"><strong>28.3.1: </strong></span><span class="bold"><strong> What are server SQL modes? </strong></span></p><p> Server SQL modes define what SQL syntax MySQL should support and what kind of data validation checks it should perform. This makes it easier to use MySQL in different environments and to use MySQL together with other database servers. The MySQL Server apply these modes individually to different clients. For more information, see <a href="server-administration.html#server-sql-mode" title="5.2.6. SQL Modes">Section 5.2.6, “SQL Modes”</a>. </p><p><a name="qandaitem-28-3-2"></a><span class="bold"><strong>28.3.2: </strong></span><span class="bold"><strong> How many server SQL modes are there? </strong></span></p><p> Each mode can be independently switched on and off. See <a href="server-administration.html#server-sql-mode" title="5.2.6. SQL Modes">Section 5.2.6, “SQL Modes”</a>, for a complete list of available modes. </p><p><a name="qandaitem-28-3-3"></a><span class="bold"><strong>28.3.3: </strong></span><span class="bold"><strong> How do you determine the server SQL mode? </strong></span></p><p> You can set the default SQL mode (for <span><strong class="command">mysqld</strong></span> startup) with the <code class="option">--sql-mode</code> option. Using the statement <code class="literal">SET [SESSION|GLOBAL] sql_mode='<em class="replaceable"><code>modes</code></em>'</code>, you can change the settings from within a connection, either locally to the connection, or to take effect globally. You can retrieve the current mode by issuing a <code class="literal">SELECT @@sql_mode</code> statement. </p><p><a name="qandaitem-28-3-4"></a><span class="bold"><strong>28.3.4: </strong></span><span class="bold"><strong> Is the mode dependent on the database or connection? </strong></span></p><p> A mode is not linked to a particular database. Modes can be set locally to the session (connection), or globally for the server. you can change these settings using <code class="literal">SET [SESSION|GLOBAL] sql_mode='<em class="replaceable"><code>modes</code></em>'</code>. </p><p><a name="qandaitem-28-3-5"></a><span class="bold"><strong>28.3.5: </strong></span><span class="bold"><strong> Can the rules for strict mode be extended? </strong></span></p><p> When we refer to <span class="emphasis"><em>strict mode</em></span>, we mean a mode where at least one of the modes <code class="literal">TRADITIONAL</code>, <code class="literal">STRICT_TRANS_TABLES</code>, or <code class="literal">STRICT_ALL_TABLES</code> is enabled. Options can be combined, so you can add additional restrictions to a mode. See <a href="server-administration.html#server-sql-mode" title="5.2.6. SQL Modes">Section 5.2.6, “SQL Modes”</a>, for more information. </p><p><a name="qandaitem-28-3-6"></a><span class="bold"><strong>28.3.6: </strong></span><span class="bold"><strong> Does strict mode impact performance? </strong></span></p><p> The intensive validation of input data that some settings requires more time than if the validation is not done. While the performance impact is not that great, if you do not require such validation (perhaps your application already handles all of this), then MySQL gives you the option of leaving strict mode disabled. However — if you do require it — strict mode can provide such validation. </p><p><a name="qandaitem-28-3-7"></a><span class="bold"><strong>28.3.7: </strong></span><span class="bold"><strong> What is the default server SQL mode when My SQL 5.0 is installed? </strong></span></p><p> By default, no special modes are enabled. See <a href="server-administration.html#server-sql-mode" title="5.2.6. SQL Modes">Section 5.2.6, “SQL Modes”</a>, for information about all available modes and MySQL's default behavior. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-stored-procs"></a>A.4. MySQL 5.0 FAQ — Stored Procedures</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-4-1">28.4.1: </a> Does MySQL 5.0 support stored procedures? </p></li><li><p><a href="faqs.html#qandaitem-28-4-2">28.4.2: </a> Where can I find documentation for MySQL stored procedures and stored functions? </p></li><li><p><a href="faqs.html#qandaitem-28-4-3">28.4.3: </a> Is there a discussion forum for MySQL stored procedures? </p></li><li><p><a href="faqs.html#qandaitem-28-4-4">28.4.4: </a> Where can I find the ANSI SQL 2003 specification for stored procedures? </p></li><li><p><a href="faqs.html#qandaitem-28-4-5">28.4.5: </a> How do you manage stored routines? </p></li><li><p><a href="faqs.html#qandaitem-28-4-6">28.4.6: </a> Is there a way to view all stored procedures and stored functions in a given database? </p></li><li><p><a href="faqs.html#qandaitem-28-4-7">28.4.7: </a> Where are stored procedures stored? </p></li><li><p><a href="faqs.html#qandaitem-28-4-8">28.4.8: </a> Is it possible to group stored procedures or stored functions into packages? </p></li><li><p><a href="faqs.html#qandaitem-28-4-9">28.4.9: </a> Can a stored procedure call another stored procedure? </p></li><li><p><a href="faqs.html#qandaitem-28-4-10">28.4.10: </a> Can a stored procedure call a trigger? </p></li><li><p><a href="faqs.html#qandaitem-28-4-11">28.4.11: </a> Can a stored procedure access tables? </p></li><li><p><a href="faqs.html#qandaitem-28-4-12">28.4.12: </a> Do stored procedures have a statement for raising application errors? </p></li><li><p><a href="faqs.html#qandaitem-28-4-13">28.4.13: </a> Do stored procedures provide exception handling? </p></li><li><p><a href="faqs.html#qandaitem-28-4-14">28.4.14: </a> Can MySQL 5.0 stored routines return result sets? </p></li><li><p><a href="faqs.html#qandaitem-28-4-15">28.4.15: </a> Is <code class="literal">WITH RECOMPILE</code> supported for stored procedures? </p></li><li><p><a href="faqs.html#qandaitem-28-4-16">28.4.16: </a> Is there a MySQL equivalent to using <code class="literal">mod_plsql</code> as a gateway on Apache to talk directly to a stored procedure in the database? </p></li><li><p><a href="faqs.html#qandaitem-28-4-17">28.4.17: </a> Can I pass an array as input to a stored procedure? </p></li><li><p><a href="faqs.html#qandaitem-28-4-18">28.4.18: </a> Can I pass a cursor as an <code class="literal">IN</code> parameter to a stored procedure? </p></li><li><p><a href="faqs.html#qandaitem-28-4-19">28.4.19: </a> Can I return a cursor as an <code class="literal">OUT</code> parameter from a stored procedure? </p></li><li><p><a href="faqs.html#qandaitem-28-4-20">28.4.20: </a> Can I print out a variable's value within a stored routine for debugging purposes? </p></li><li><p><a href="faqs.html#qandaitem-28-4-21">28.4.21: </a> Can I commit or roll back transactions inside a stored procedure? </p></li><li><p><a href="faqs.html#qandaitem-28-4-22">28.4.22: </a> Do MySQL 5.0 stored procedures and functions work with replication? </p></li><li><p><a href="faqs.html#qandaitem-28-4-23">28.4.23: </a> Are stored procedures and functions created on a master server replicated to a slave? </p></li><li><p><a href="faqs.html#qandaitem-28-4-24">28.4.24: </a> How are actions that take place inside stored procedures and functions replicated? </p></li><li><p><a href="faqs.html#qandaitem-28-4-25">28.4.25: </a> Are there special security requirements for using stored procedures and functions together with replication? </p></li><li><p><a href="faqs.html#qandaitem-28-4-26">28.4.26: </a> What limitations exist for replicating stored procedure and function actions? </p></li><li><p><a href="faqs.html#qandaitem-28-4-27">28.4.27: </a> Do the preceding limitations affect MySQL's ability to do point-in-time recovery? </p></li><li><p><a href="faqs.html#qandaitem-28-4-28">28.4.28: </a> What is being done to correct the aforementioned limitations? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-4-1"></a><span class="bold"><strong>28.4.1: </strong></span><span class="bold"><strong> Does MySQL 5.0 support stored procedures? </strong></span></p><p> Yes. MySQL 5.0 supports two types of stored routines — stored procedures and stored functions. </p><p><a name="qandaitem-28-4-2"></a><span class="bold"><strong>28.4.2: </strong></span><span class="bold"><strong> Where can I find documentation for MySQL stored procedures and stored functions? </strong></span></p><p> See <a href="stored-procedures.html" title="Chapter 18. Stored Procedures and Functions">Chapter 18, <i>Stored Procedures and Functions</i></a>. </p><p><a name="qandaitem-28-4-3"></a><span class="bold"><strong>28.4.3: </strong></span><span class="bold"><strong> Is there a discussion forum for MySQL stored procedures? </strong></span></p><p> Yes. See <a href="http://forums.mysql.com/list.php?98" target="_top">http://forums.mysql.com/list.php?98</a>. </p><p><a name="qandaitem-28-4-4"></a><span class="bold"><strong>28.4.4: </strong></span><span class="bold"><strong> Where can I find the ANSI SQL 2003 specification for stored procedures? </strong></span></p><p> Unfortunately, the official specifications are not freely available (ANSI makes them available for purchase). However, there are books — such as <em class="citetitle">SQL-99 Complete, Really</em> by Peter Gulutzan and Trudy Pelzer — which give a comprehensive overview of the standard, including coverage of stored procedures. </p><p><a name="qandaitem-28-4-5"></a><span class="bold"><strong>28.4.5: </strong></span><span class="bold"><strong> How do you manage stored routines? </strong></span></p><p> It is always good practice to use a clear naming scheme for your stored routines. You can manage stored procedures with <code class="literal">CREATE [FUNCTION|PROCEDURE]</code>, <code class="literal">ALTER [FUNCTION|PROCEDURE]</code>, <code class="literal">DROP [FUNCTION|PROCEDURE]</code>, and <code class="literal">SHOW CREATE [FUNCTION|PROCEDURE]</code>. You can obtain information about existing stored procedures using the <code class="literal">ROUTINES</code> table in the <code class="literal">INFORMATION_SCHEMA</code> database (see <a href="information-schema.html#routines-table" title="21.14. The INFORMATION_SCHEMA ROUTINES Table">Section 21.14, “The <code class="literal">INFORMATION_SCHEMA ROUTINES</code> Table”</a>). </p><p><a name="qandaitem-28-4-6"></a><span class="bold"><strong>28.4.6: </strong></span><span class="bold"><strong> Is there a way to view all stored procedures and stored functions in a given database? </strong></span></p><p> Yes. For a database named <em class="replaceable"><code>dbname</code></em>, use this query on the <code class="literal">INFORMATION_SCHEMA.ROUTINES</code> table: </p><pre class="programlisting">SELECT ROUTINE_TYPE, ROUTINE_NAME FROM INFORMATION_SCHEMA.ROUTINES WHERE ROUTINE_SCHEMA='<em class="replaceable"><code>dbname</code></em>'; </pre><p> For more information, see <a href="information-schema.html#routines-table" title="21.14. The INFORMATION_SCHEMA ROUTINES Table">Section 21.14, “The <code class="literal">INFORMATION_SCHEMA ROUTINES</code> Table”</a>. </p><p> The body of a stored routine can be viewed using <code class="literal">SHOW CREATE FUNCTION</code> (for a stored function) or <code class="literal">SHOW CREATE PROCEDURE</code> (for a stored procedure). See <a href="sql-syntax.html#show-create-procedure" title="12.5.4.5. SHOW CREATE PROCEDURE and SHOW CREATE FUNCTION Syntax">Section 12.5.4.5, “<code class="literal">SHOW CREATE PROCEDURE</code> and <code class="literal">SHOW CREATE FUNCTION</code> Syntax”</a>, for more information. </p><p><a name="qandaitem-28-4-7"></a><span class="bold"><strong>28.4.7: </strong></span><span class="bold"><strong> Where are stored procedures stored? </strong></span></p><p> In the <code class="literal">proc</code> table of the <code class="literal">mysql</code> system database. However, you should not access the tables in the system database directly. Instead, use <code class="literal">SHOW CREATE FUNCTION</code> to obtain information about stored functions, and <code class="literal">SHOW CREATE PROCEDURE</code> to obtain information about stored procedures. See <a href="sql-syntax.html#show-create-procedure" title="12.5.4.5. SHOW CREATE PROCEDURE and SHOW CREATE FUNCTION Syntax">Section 12.5.4.5, “<code class="literal">SHOW CREATE PROCEDURE</code> and <code class="literal">SHOW CREATE FUNCTION</code> Syntax”</a>, for more information about these statements. </p><p> You can also query the <code class="literal">ROUTINES</code> table in the <code class="literal">INFORMATION_SCHEMA</code> database — see <a href="information-schema.html#routines-table" title="21.14. The INFORMATION_SCHEMA ROUTINES Table">Section 21.14, “The <code class="literal">INFORMATION_SCHEMA ROUTINES</code> Table”</a>, for information about this table. </p><p><a name="qandaitem-28-4-8"></a><span class="bold"><strong>28.4.8: </strong></span><span class="bold"><strong> Is it possible to group stored procedures or stored functions into packages? </strong></span></p><p> No. This is not supported in MySQL 5.0. </p><p><a name="qandaitem-28-4-9"></a><span class="bold"><strong>28.4.9: </strong></span><span class="bold"><strong> Can a stored procedure call another stored procedure? </strong></span></p><p> Yes. </p><p><a name="qandaitem-28-4-10"></a><span class="bold"><strong>28.4.10: </strong></span><span class="bold"><strong> Can a stored procedure call a trigger? </strong></span></p><p> A stored procedure can execute an SQL statement, such as an <code class="literal">UPDATE</code>, that causes a trigger to fire. </p><p><a name="qandaitem-28-4-11"></a><span class="bold"><strong>28.4.11: </strong></span><span class="bold"><strong> Can a stored procedure access tables? </strong></span></p><p> Yes. A stored procedure can access one or more tables as required. </p><p><a name="qandaitem-28-4-12"></a><span class="bold"><strong>28.4.12: </strong></span><span class="bold"><strong> Do stored procedures have a statement for raising application errors? </strong></span></p><p> Not in MySQL 5.0. We intend to implement the SQL standard <code class="literal">SIGNAL</code> and <code class="literal">RESIGNAL</code> statements in a future MySQL release. </p><p><a name="qandaitem-28-4-13"></a><span class="bold"><strong>28.4.13: </strong></span><span class="bold"><strong> Do stored procedures provide exception handling? </strong></span></p><p> MySQL implements <code class="literal">HANDLER</code> definitions according to the SQL standard. See <a href="stored-procedures.html#declare-handlers" title="18.2.8.2. DECLARE Handlers">Section 18.2.8.2, “<code class="literal">DECLARE</code> Handlers”</a>, for details. </p><p><a name="qandaitem-28-4-14"></a><span class="bold"><strong>28.4.14: </strong></span><span class="bold"><strong> Can MySQL 5.0 stored routines return result sets? </strong></span></p><p> <span class="emphasis"><em>Stored procedures</em></span> can, but stored functions cannot. If you perform an ordinary <code class="literal">SELECT</code> inside a stored procedure, the result set is returned directly to the client. You need to use the MySQL 4.1 (or above) client-server protocol for this to work. This means that — for instance — in PHP, you need to use the <code class="literal">mysqli</code> extension rather than the old <code class="literal">mysql</code> extension. </p><p><a name="qandaitem-28-4-15"></a><span class="bold"><strong>28.4.15: </strong></span><span class="bold"><strong> Is <code class="literal">WITH RECOMPILE</code> supported for stored procedures? </strong></span></p><p> Not in MySQL 5.0. </p><p><a name="qandaitem-28-4-16"></a><span class="bold"><strong>28.4.16: </strong></span><span class="bold"><strong> Is there a MySQL equivalent to using <code class="literal">mod_plsql</code> as a gateway on Apache to talk directly to a stored procedure in the database? </strong></span></p><p> There is no equivalent in MySQL 5.0. </p><p><a name="qandaitem-28-4-17"></a><span class="bold"><strong>28.4.17: </strong></span><span class="bold"><strong> Can I pass an array as input to a stored procedure? </strong></span></p><p> Not in MySQL 5.0. </p><p><a name="qandaitem-28-4-18"></a><span class="bold"><strong>28.4.18: </strong></span><span class="bold"><strong> Can I pass a cursor as an <code class="literal">IN</code> parameter to a stored procedure? </strong></span></p><p> In MySQL 5.0, cursors are available inside stored procedures only. </p><p><a name="qandaitem-28-4-19"></a><span class="bold"><strong>28.4.19: </strong></span><span class="bold"><strong> Can I return a cursor as an <code class="literal">OUT</code> parameter from a stored procedure? </strong></span></p><p> In MySQL 5.0, cursors are available inside stored procedures only. However, if you do not open a cursor on a <code class="literal">SELECT</code>, the result will be sent directly to the client. You can also <code class="literal">SELECT INTO</code> variables. See <a href="sql-syntax.html#select" title="12.2.7. SELECT Syntax">Section 12.2.7, “<code class="literal">SELECT</code> Syntax”</a>. </p><p><a name="qandaitem-28-4-20"></a><span class="bold"><strong>28.4.20: </strong></span><span class="bold"><strong> Can I print out a variable's value within a stored routine for debugging purposes? </strong></span></p><p> Yes, you can do this in a <span class="emphasis"><em>stored procedure</em></span>, but not in a stored function. If you perform an ordinary <code class="literal">SELECT</code> inside a stored procedure, the result set is returned directly to the client. You will need to use the MySQL 4.1 (or above) client-server protocol for this to work. This means that — for instance — in PHP, you need to use the <code class="literal">mysqli</code> extension rather than the old <code class="literal">mysql</code> extension. </p><p><a name="qandaitem-28-4-21"></a><span class="bold"><strong>28.4.21: </strong></span><span class="bold"><strong> Can I commit or roll back transactions inside a stored procedure? </strong></span></p><p> Yes. However, you cannot perform transactional operations within a stored function. </p><p><a name="qandaitem-28-4-22"></a><span class="bold"><strong>28.4.22: </strong></span><span class="bold"><strong> Do MySQL 5.0 stored procedures and functions work with replication? </strong></span></p><p> Yes, standard actions carried out in stored procedures and functions are replicated from a master MySQL server to a slave server. There are a few limitations that are described in detail in <a href="stored-procedures.html#stored-procedure-logging" title="18.4. Binary Logging of Stored Routines and Triggers">Section 18.4, “Binary Logging of Stored Routines and Triggers”</a>. </p><p><a name="qandaitem-28-4-23"></a><span class="bold"><strong>28.4.23: </strong></span><span class="bold"><strong> Are stored procedures and functions created on a master server replicated to a slave? </strong></span></p><p> Yes, creation of stored procedures and functions carried out through normal DDL statements on a master server are replicated to a slave, so the objects will exist on both servers. <code class="literal">ALTER</code> and <code class="literal">DROP</code> statements for stored procedures and functions are also replicated. </p><p><a name="qandaitem-28-4-24"></a><span class="bold"><strong>28.4.24: </strong></span><span class="bold"><strong> How are actions that take place inside stored procedures and functions replicated? </strong></span></p><p> MySQL records each DML event that occurs in a stored procedure and replicates those individual actions to a slave server. The actual calls made to execute stored procedures are not replicated. </p><p> Stored functions that change data are logged as function invocations, not as the DML events that occur inside each function. </p><p><a name="qandaitem-28-4-25"></a><span class="bold"><strong>28.4.25: </strong></span><span class="bold"><strong> Are there special security requirements for using stored procedures and functions together with replication? </strong></span></p><p> Yes. Because a slave server has authority to execute any statement read from a master's binary log, special security constraints exist for using stored functions with replication. If replication or binary logging in general (for the purpose of point-in-time recovery) is active, then MySQL DBAs have two security options open to them: </p><div class="orderedlist"><ol type="1"><li><p> Any user wishing to create stored functions must be granted the <code class="literal">SUPER</code> privilege. </p></li><li><p> Alternatively, a DBA can set the <code class="literal">log_bin_trust_function_creators</code> system variable to 1, which enables anyone with the standard <code class="literal">CREATE ROUTINE</code> privilege to create stored functions. </p></li></ol></div><p> </p><p><a name="qandaitem-28-4-26"></a><span class="bold"><strong>28.4.26: </strong></span><span class="bold"><strong> What limitations exist for replicating stored procedure and function actions? </strong></span></p><p> Non-deterministic (random) or time-based actions embedded in stored procedures may not replicate properly. By their very nature, randomly produced results are not predictable and cannot be exactly reproduced, and therefore, random actions replicated to a slave will not mirror those performed on a master. Note that declaring stored functions to be <code class="literal">DETERMINISTIC</code> or setting the <code class="literal">log_bin_trust_function_creators</code> system variable to 0 will not allow random-valued operations to be invoked. </p><p> In addition, time-based actions cannot be reproduced on a slave because the timing of such actions in a stored procedure is not reproducible through the binary log used for replication. It records only DML events and does not factor in timing constraints. </p><p> Finally, non-transactional tables for which errors occur during large DML actions (such as bulk inserts) may experience replication issues in that a master may be partially updated from DML activity, but no updates are done to the slave because of the errors that occurred. A workaround is for a function's DML actions to be carried out with the <code class="literal">IGNORE</code> keyword so that updates on the master that cause errors are ignored and updates that do not cause errors are replicated to the slave. </p><p><a name="qandaitem-28-4-27"></a><span class="bold"><strong>28.4.27: </strong></span><span class="bold"><strong> Do the preceding limitations affect MySQL's ability to do point-in-time recovery? </strong></span></p><p> The same limitations that affect replication do affect point-in-time recovery. </p><p><a name="qandaitem-28-4-28"></a><span class="bold"><strong>28.4.28: </strong></span><span class="bold"><strong> What is being done to correct the aforementioned limitations? </strong></span></p><p> MySQL 5.1 implements <em class="firstterm">row-based replication</em>, which resolves the limitations mentioned earlier. </p><p> We do not plan to backport row-based replication to MySQL 5.0. For additional information, see <a href="http://dev.mysql.com/doc/refman" target="_top"><em class="citetitle">Row-Based Replication</em>, in the <em class="citetitle">MySQL 5.1 Manual</em></a>. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-triggers"></a>A.5. MySQL 5.0 FAQ — Triggers</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-5-1">28.5.1: </a> Where can I find the documentation for MySQL 5.0 triggers? </p></li><li><p><a href="faqs.html#qandaitem-28-5-2">28.5.2: </a> Is there a discussion forum for MySQL Triggers? </p></li><li><p><a href="faqs.html#qandaitem-28-5-3">28.5.3: </a> Does MySQL 5.0 have statement-level or row-level triggers? </p></li><li><p><a href="faqs.html#qandaitem-28-5-4">28.5.4: </a> Are there any default triggers? </p></li><li><p><a href="faqs.html#qandaitem-28-5-5">28.5.5: </a> How are triggers managed in MySQL? </p></li><li><p><a href="faqs.html#qandaitem-28-5-6">28.5.6: </a> Is there a way to view all triggers in a given database? </p></li><li><p><a href="faqs.html#qandaitem-28-5-7">28.5.7: </a> Where are triggers stored? </p></li><li><p><a href="faqs.html#qandaitem-28-5-8">28.5.8: </a> Can a trigger call a stored procedure? </p></li><li><p><a href="faqs.html#qandaitem-28-5-9">28.5.9: </a> Can triggers access tables? </p></li><li><p><a href="faqs.html#qandaitem-28-5-10">28.5.10: </a> Can triggers call an external application through a UDF? </p></li><li><p><a href="faqs.html#qandaitem-28-5-11">28.5.11: </a> Is possible for a trigger to update tables on a remote server? </p></li><li><p><a href="faqs.html#qandaitem-28-5-12">28.5.12: </a> Do triggers work with replication? </p></li><li><p><a href="faqs.html#qandaitem-28-5-13">28.5.13: </a> How are actions carried out through triggers on a master replicated to a slave? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-5-1"></a><span class="bold"><strong>28.5.1: </strong></span><span class="bold"><strong> Where can I find the documentation for MySQL 5.0 triggers? </strong></span></p><p> See <a href="triggers.html" title="Chapter 19. Triggers">Chapter 19, <i>Triggers</i></a>. </p><p><a name="qandaitem-28-5-2"></a><span class="bold"><strong>28.5.2: </strong></span><span class="bold"><strong> Is there a discussion forum for MySQL Triggers? </strong></span></p><p> Yes. It is available at <a href="http://forums.mysql.com/list.php?99" target="_top">http://forums.mysql.com/list.php?99</a>. </p><p><a name="qandaitem-28-5-3"></a><span class="bold"><strong>28.5.3: </strong></span><span class="bold"><strong> Does MySQL 5.0 have statement-level or row-level triggers? </strong></span></p><p> In MySQL 5.0, all triggers are <code class="literal">FOR EACH ROW</code> — that is, the trigger is activated for each row that is inserted, updated, or deleted. MySQL 5.0 does not support triggers using <code class="literal">FOR EACH STATEMENT</code>. </p><p><a name="qandaitem-28-5-4"></a><span class="bold"><strong>28.5.4: </strong></span><span class="bold"><strong> Are there any default triggers? </strong></span></p><p> Not explicitly. MySQL does have specific special behavior for some <code class="literal">TIMESTAMP</code> columns, as well as for columns which are defined using <code class="literal">AUTO_INCREMENT</code>. </p><p><a name="qandaitem-28-5-5"></a><span class="bold"><strong>28.5.5: </strong></span><span class="bold"><strong> How are triggers managed in MySQL? </strong></span></p><p> In MySQL 5.0, triggers can be created using the <code class="literal">CREATE TRIGGER</code> statement, and dropped using <code class="literal">DROP TRIGGER</code>. See <a href="triggers.html#create-trigger" title="19.1. CREATE TRIGGER Syntax">Section 19.1, “<code class="literal">CREATE TRIGGER</code> Syntax”</a>, and <a href="triggers.html#drop-trigger" title="19.2. DROP TRIGGER Syntax">Section 19.2, “<code class="literal">DROP TRIGGER</code> Syntax”</a>, for more about these statements. </p><p> Information about triggers can be obtained by querying the <code class="literal">INFORMATION_SCHEMA.TRIGGERS</code> table. See <a href="information-schema.html#triggers-table" title="21.16. The INFORMATION_SCHEMA TRIGGERS Table">Section 21.16, “The <code class="literal">INFORMATION_SCHEMA TRIGGERS</code> Table”</a>. </p><p><a name="qandaitem-28-5-6"></a><span class="bold"><strong>28.5.6: </strong></span><span class="bold"><strong> Is there a way to view all triggers in a given database? </strong></span></p><p> Yes. You can obtain a listing of all triggers defined on database <code class="literal">dbname</code> using a query on the <code class="literal">INFORMATION_SCHEMA.TRIGGERS</code> table such as the one shown here: </p><pre class="programlisting">SELECT TRIGGER_NAME, EVENT_MANIPULATION, EVENT_OBJECT_TABLE, ACTION_STATEMENT FROM INFORMATION_SCHEMA.TRIGGERS WHERE TRIGGER_SCHEMA='<em class="replaceable"><code>dbname</code></em>'; </pre><p> For more information about this table, see <a href="information-schema.html#triggers-table" title="21.16. The INFORMATION_SCHEMA TRIGGERS Table">Section 21.16, “The <code class="literal">INFORMATION_SCHEMA TRIGGERS</code> Table”</a>. </p><p> You can also use the <code class="literal">SHOW TRIGGERS</code> statement, which is specific to MySQL. See <a href="sql-syntax.html#show-triggers" title="12.5.4.26. SHOW TRIGGERS Syntax">Section 12.5.4.26, “<code class="literal">SHOW TRIGGERS</code> Syntax”</a>. </p><p><a name="qandaitem-28-5-7"></a><span class="bold"><strong>28.5.7: </strong></span><span class="bold"><strong> Where are triggers stored? </strong></span></p><p> Triggers for a table are currently stored in <code class="filename">.TRG</code> files, with one such file one per table. </p><p><a name="qandaitem-28-5-8"></a><span class="bold"><strong>28.5.8: </strong></span><span class="bold"><strong> Can a trigger call a stored procedure? </strong></span></p><p> Yes. </p><p><a name="qandaitem-28-5-9"></a><span class="bold"><strong>28.5.9: </strong></span><span class="bold"><strong> Can triggers access tables? </strong></span></p><p> A trigger can access both old and new data in its own table. Through a stored procedure, or a multiple-table update or delete statement, a trigger can also affect other tables. </p><p><a name="qandaitem-28-5-10"></a><span class="bold"><strong>28.5.10: </strong></span><span class="bold"><strong> Can triggers call an external application through a UDF? </strong></span></p><p> No, not at present. </p><p><a name="qandaitem-28-5-11"></a><span class="bold"><strong>28.5.11: </strong></span><span class="bold"><strong> Is possible for a trigger to update tables on a remote server? </strong></span></p><p> Yes. A table on a remote server could be updated using the <code class="literal">FEDERATED</code> storage engine. (See <a href="storage-engines.html#federated-storage-engine" title="13.7. The FEDERATED Storage Engine">Section 13.7, “The <code class="literal">FEDERATED</code> Storage Engine”</a>). </p><p><a name="qandaitem-28-5-12"></a><span class="bold"><strong>28.5.12: </strong></span><span class="bold"><strong> Do triggers work with replication? </strong></span></p><p> Triggers and replication in MySQL 5.0 work in the same wasy as in most other database engines: Actions carried out through triggers on a master are not replicated to a slave server. Instead, triggers that exist on tables that reside on a MySQL master server need to be created on the corresponding tables on any MySQL slave servers so that the triggers activate on the slaves as well as the master. </p><p><a name="qandaitem-28-5-13"></a><span class="bold"><strong>28.5.13: </strong></span><span class="bold"><strong> How are actions carried out through triggers on a master replicated to a slave? </strong></span></p><p> First, the triggers that exist on a master must be re-created on the slave server. Once this is done, the replication flow works as any other standard DML statement that participates in replication. For example, consider a table <code class="literal">EMP</code> that has an <code class="literal">AFTER</code> insert trigger, which exists on a master MySQL server. The same <code class="literal">EMP</code> table and <code class="literal">AFTER</code> insert trigger exist on the slave server as well. The replication flow would be: </p><div class="orderedlist"><ol type="1"><li><p> An <code class="literal">INSERT</code> statement is made to <code class="literal">EMP</code>. </p></li><li><p> The <code class="literal">AFTER</code> trigger on <code class="literal">EMP</code> activates. </p></li><li><p> The <code class="literal">INSERT</code> statement is written to the binary log. </p></li><li><p> The replication slave picks up the <code class="literal">INSERT</code> statement to <code class="literal">EMP</code> and executes it. </p></li><li><p> The <code class="literal">AFTER</code> trigger on <code class="literal">EMP</code> that exists on the slave activates. </p></li></ol></div></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-views"></a>A.6. MySQL 5.0 FAQ — Views</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-6-1">28.6.1: </a> Where can I find documentation covering MySQL Views? </p></li><li><p><a href="faqs.html#qandaitem-28-6-2">28.6.2: </a> Is there a discussion forum for MySQL Views? </p></li><li><p><a href="faqs.html#qandaitem-28-6-3">28.6.3: </a> What happens to a view if an underlying table is dropped or renamed? </p></li><li><p><a href="faqs.html#qandaitem-28-6-4">28.6.4: </a> Does MySQL 5.0 have table snapshots? </p></li><li><p><a href="faqs.html#qandaitem-28-6-5">28.6.5: </a> Does MySQL 5.0 have materialized views? </p></li><li><p><a href="faqs.html#qandaitem-28-6-6">28.6.6: </a> Can you insert into views that are based on joins? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-6-1"></a><span class="bold"><strong>28.6.1: </strong></span><span class="bold"><strong> Where can I find documentation covering MySQL Views? </strong></span></p><p> See <a href="views.html" title="Chapter 20. Views">Chapter 20, <i>Views</i></a>. </p><p><a name="qandaitem-28-6-2"></a><span class="bold"><strong>28.6.2: </strong></span><span class="bold"><strong> Is there a discussion forum for MySQL Views? </strong></span></p><p> Yes. See <a href="http://forums.mysql.com/list.php?100" target="_top">http://forums.mysql.com/list.php?100</a> </p><p><a name="qandaitem-28-6-3"></a><span class="bold"><strong>28.6.3: </strong></span><span class="bold"><strong> What happens to a view if an underlying table is dropped or renamed? </strong></span></p><p> After a view has been created, it is possible to drop or alter a table or view to which the definition refers. To check a view definition for problems of this kind, use the <code class="literal">CHECK TABLE</code> statement. (See <a href="sql-syntax.html#check-table" title="12.5.2.3. CHECK TABLE Syntax">Section 12.5.2.3, “<code class="literal">CHECK TABLE</code> Syntax”</a>.) </p><p><a name="qandaitem-28-6-4"></a><span class="bold"><strong>28.6.4: </strong></span><span class="bold"><strong> Does MySQL 5.0 have table snapshots? </strong></span></p><p> No. </p><p><a name="qandaitem-28-6-5"></a><span class="bold"><strong>28.6.5: </strong></span><span class="bold"><strong> Does MySQL 5.0 have materialized views? </strong></span></p><p> No. </p><p><a name="qandaitem-28-6-6"></a><span class="bold"><strong>28.6.6: </strong></span><span class="bold"><strong> Can you insert into views that are based on joins? </strong></span></p><p> It is possible, provided that your <code class="literal">INSERT</code> statement has a column list that makes it clear there's only one table involved. </p><p> You <span class="emphasis"><em>cannot</em></span> insert into multiple tables with a single insert on a view. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-information-schema"></a>A.7. MySQL 5.0 FAQ — <code class="literal">INFORMATION_SCHEMA</code></h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-7-1">28.7.1: </a> Where can I find documentation for the MySQL <code class="literal">INFORMATION_SCHEMA</code> database? </p></li><li><p><a href="faqs.html#qandaitem-28-7-2">28.7.2: </a> Is there a discussion forum for <code class="literal">INFORMATION_SCHEMA</code>? </p></li><li><p><a href="faqs.html#qandaitem-28-7-3">28.7.3: </a> Where can I find the ANSI SQL 2003 specification for <code class="literal">INFORMATION_SCHEMA</code>? </p></li><li><p><a href="faqs.html#qandaitem-28-7-4">28.7.4: </a> What is the difference between the Oracle Data Dictionary and MySQL's <code class="literal">INFORMATION_SCHEMA</code>? </p></li><li><p><a href="faqs.html#qandaitem-28-7-5">28.7.5: </a> Can I add to or otherwise modify the tables found in the <code class="literal">INFORMATION_SCHEMA</code> database? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-7-1"></a><span class="bold"><strong>28.7.1: </strong></span><span class="bold"><strong> Where can I find documentation for the MySQL <code class="literal">INFORMATION_SCHEMA</code> database? </strong></span></p><p> See <a href="information-schema.html" title="Chapter 21. INFORMATION_SCHEMA Tables">Chapter 21, <i><code class="literal">INFORMATION_SCHEMA</code> Tables</i></a> </p><p><a name="qandaitem-28-7-2"></a><span class="bold"><strong>28.7.2: </strong></span><span class="bold"><strong> Is there a discussion forum for <code class="literal">INFORMATION_SCHEMA</code>? </strong></span></p><p> See <a href="http://forums.mysql.com/list.php?101" target="_top">http://forums.mysql.com/list.php?101</a>. </p><p><a name="qandaitem-28-7-3"></a><span class="bold"><strong>28.7.3: </strong></span><span class="bold"><strong> Where can I find the ANSI SQL 2003 specification for <code class="literal">INFORMATION_SCHEMA</code>? </strong></span></p><p> Unfortunately, the official specifications are not freely available. (ANSI makes them available for purchase.) However, there are books available — such as <em class="citetitle">SQL-99 Complete, Really</em> by Peter Gulutzan and Trudy Pelzer — which give a comprehensive overview of the standard, including <code class="literal">INFORMATION_SCHEMA</code>. </p><p><a name="qandaitem-28-7-4"></a><span class="bold"><strong>28.7.4: </strong></span><span class="bold"><strong> What is the difference between the Oracle Data Dictionary and MySQL's <code class="literal">INFORMATION_SCHEMA</code>? </strong></span></p><p> Both Oracle and MySQL provide metadata in tables. However, Oracle and MySQL use different table names and column names. MySQL's implementation is more similar to those found in DB2 and SQL Server, which also support <code class="literal">INFORMATION_SCHEMA</code> as defined in the SQL standard. </p><p><a name="qandaitem-28-7-5"></a><span class="bold"><strong>28.7.5: </strong></span><span class="bold"><strong> Can I add to or otherwise modify the tables found in the <code class="literal">INFORMATION_SCHEMA</code> database? </strong></span></p><p> No. Since applications may rely on a certain standard structure, this should not be modified. For this reason, <span class="emphasis"><em>MySQL AB cannot support bugs or other issues which result from modifying <code class="literal">INFORMATION_SCHEMA</code> tables or data</em></span>. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-migration"></a>A.8. MySQL 5.0 FAQ — Migration</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-8-1">28.8.1: </a> Where can I find information on how to migrate from MySQL 4.1 to MySQL 5.0? </p></li><li><p><a href="faqs.html#qandaitem-28-8-2">28.8.2: </a> How has storage engine (table type) support changed in MySQL 5.0 from previous versions? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-8-1"></a><span class="bold"><strong>28.8.1: </strong></span><span class="bold"><strong> Where can I find information on how to migrate from MySQL 4.1 to MySQL 5.0? </strong></span></p><p> For detailed upgrade information, see <a href="installing.html#upgrade" title="2.4.17. Upgrading MySQL">Section 2.4.17, “Upgrading MySQL”</a>. We recommend that you do not skip a major version when upgrading, but rather complete the process in steps, upgrading from one major version to the next in each step. This may seem more complicated, but it will you save time and trouble — if you encounter problems during the upgrade, their origin will be easier to identify, either by you or — if you have a MySQL Enterprise subscription — by MySQL support. </p><p><a name="qandaitem-28-8-2"></a><span class="bold"><strong>28.8.2: </strong></span><span class="bold"><strong> How has storage engine (table type) support changed in MySQL 5.0 from previous versions? </strong></span></p><p> Storage engine support has changed as follows: </p><div class="itemizedlist"><ul type="disc"><li><p> Support for <code class="literal">ISAM</code> tables was removed in MySQL 5.0 and you should now use the <code class="literal">MyISAM</code> storage engine in place of <code class="literal">ISAM</code>. To convert a table <em class="replaceable"><code>tblname</code></em> from <code class="literal">ISAM</code> to <code class="literal">MyISAM</code>, simply issue a statement such as this one: </p><pre class="programlisting">ALTER TABLE <em class="replaceable"><code>tblname</code></em> ENGINE=MYISAM;</pre></li><li><p> Internal <code class="literal">RAID</code> for <code class="literal">MyISAM</code> tables was also removed in MySQL 5.0. This was formerly used to allow large tables in file systems that did not support file sizes greater than 2GB. All modern file systems allow for larger tables; in addition, there are now other solutions such as <code class="literal">MERGE</code> tables and views. </p></li><li><p> The <code class="literal">VARCHAR</code> column type now retains trailing spaces in all storage engines. </p></li><li><p> <code class="literal">MEMORY</code> tables (formerly known as <code class="literal">HEAP</code> tables) can also contain <code class="literal">VARCHAR</code> columns. </p></li></ul></div><p> </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-security"></a>A.9. MySQL 5.0 FAQ — Security</h2></div></div></div><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-9-1">28.9.1: </a> Where can I find documentation that addresses security issues for MySQL? </p></li><li><p><a href="faqs.html#qandaitem-28-9-2">28.9.2: </a> Does MySQL 5.0 have native support for SSL? </p></li><li><p><a href="faqs.html#qandaitem-28-9-3">28.9.3: </a> Is SSL support be built into MySQL binaries, or must I recompile the binary myself to enable it? </p></li><li><p><a href="faqs.html#qandaitem-28-9-4">28.9.4: </a> Does MySQL 5.0 have built-in authentication against LDAP directories? </p></li><li><p><a href="faqs.html#qandaitem-28-9-5">28.9.5: </a> Does MySQL 5.0 include support for Roles Based Access Control (RBAC)? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-9-1"></a><span class="bold"><strong>28.9.1: </strong></span><span class="bold"><strong> Where can I find documentation that addresses security issues for MySQL? </strong></span></p><p> The best place to start is <a href="server-administration.html#security" title="5.6. General Security Issues">Section 5.6, “General Security Issues”</a>. </p><p> Other portions of the MySQL Documentation which you may find useful with regard to specific security concerns include the following: </p><div class="itemizedlist"><ul type="disc"><li><p> <a href="server-administration.html#security-guidelines" title="5.6.1. General Security Guidelines">Section 5.6.1, “General Security Guidelines”</a>. </p></li><li><p> <a href="server-administration.html#security-against-attack" title="5.6.2. Making MySQL Secure Against Attackers">Section 5.6.2, “Making MySQL Secure Against Attackers”</a>. </p></li><li><p> <a href="error-handling.html#resetting-permissions" title="B.1.4.1. How to Reset the Root Password">Section B.1.4.1, “How to Reset the Root Password”</a>. </p></li><li><p> <a href="server-administration.html#changing-mysql-user" title="5.6.5. How to Run MySQL as a Normal User">Section 5.6.5, “How to Run MySQL as a Normal User”</a>. </p></li><li><p> <a href="extending-mysql.html#udf-security" title="26.2.4.6. User-Defined Function Security Precautions">Section 26.2.4.6, “User-Defined Function Security Precautions”</a>. </p></li><li><p> <a href="server-administration.html#privileges-options" title="5.6.3. Security-Related mysqld Options">Section 5.6.3, “Security-Related <span><strong class="command">mysqld</strong></span> Options”</a>. </p></li><li><p> <a href="server-administration.html#load-data-local" title="5.6.4. Security Issues with LOAD DATA LOCAL">Section 5.6.4, “Security Issues with <code class="literal">LOAD DATA LOCAL</code>”</a>. </p></li><li><p> <a href="installing.html#post-installation" title="2.4.16. Post-Installation Setup and Testing">Section 2.4.16, “Post-Installation Setup and Testing”</a>. </p></li><li><p> <a href="installing.html#selinux" title="2.4.19.1.11. SELinux Notes">Section 2.4.19.1.11, “SELinux Notes”</a>. </p></li><li><p> <a href="server-administration.html#secure-basics" title="5.8.7.1. Basic SSL Concepts">Section 5.8.7.1, “Basic SSL Concepts”</a>. </p></li></ul></div><p> </p><p class="mnmas"><b>MySQL Enterprise</b> The MySQL Enterprise Monitor enforces best practices for maximizing the security of your servers. For more information see <a href="http://www.mysql.com/products/enterprise/advisors.html" target="_top">http://www.mysql.com/products/enterprise/advisors.html</a>. </p><p><a name="qandaitem-28-9-2"></a><span class="bold"><strong>28.9.2: </strong></span><span class="bold"><strong> Does MySQL 5.0 have native support for SSL? </strong></span></p><p> Most 5.0 binaries have support for SSL connections between the client and server. We can't currently build with the new YaSSL library everywhere, as it's still quite new and does not compile on all platforms yet. See <a href="server-administration.html#secure-connections" title="5.8.7. Using Secure Connections">Section 5.8.7, “Using Secure Connections”</a>. </p><p> You can also tunnel a connection via SSH, if (for instance) if the client application doesn't support SSL connections. For an example, see <a href="server-administration.html#windows-and-ssh" title="5.8.7.5. Connecting to MySQL Remotely from Windows with SSH">Section 5.8.7.5, “Connecting to MySQL Remotely from Windows with SSH”</a>. </p><p><a name="qandaitem-28-9-3"></a><span class="bold"><strong>28.9.3: </strong></span><span class="bold"><strong> Is SSL support be built into MySQL binaries, or must I recompile the binary myself to enable it? </strong></span></p><p> Most 5.0 binaries have SSL enabled for client-server connections that are secured, authenticated, or both. However, the YaSSL library currently does not compile on all platforms. See <a href="server-administration.html#secure-connections" title="5.8.7. Using Secure Connections">Section 5.8.7, “Using Secure Connections”</a>, for a complete listing of supported and unsupported platforms. </p><p><a name="qandaitem-28-9-4"></a><span class="bold"><strong>28.9.4: </strong></span><span class="bold"><strong> Does MySQL 5.0 have built-in authentication against LDAP directories? </strong></span></p><p> No. Support for external authentication methods is on the MySQL roadmap as a “<span class="quote">rolling feature</span>”, which means that we plan to implement it in the future, but we have not yet determined when this will be done. </p><p><a name="qandaitem-28-9-5"></a><span class="bold"><strong>28.9.5: </strong></span><span class="bold"><strong> Does MySQL 5.0 include support for Roles Based Access Control (RBAC)? </strong></span></p><p> No. Support for roles is on the MySQL roadmap as a “<span class="quote">rolling feature</span>”, which means that we plan to implement it in the future, but we have not yet determined when this will be done. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-mysql-cluster"></a>A.10. MySQL 5.0 FAQ — MySQL Cluster</h2></div></div></div><a class="indexterm" name="id3165434"></a><a class="indexterm" name="id3165446"></a><a class="indexterm" name="id3165458"></a><p> In the following section, we provide answers to questions that are most frequently asked about MySQL Cluster and the <code class="literal">NDB</code> storage engine. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-10-1">28.10.1: </a><a class="indexterm" name="id3165510"></a> What does “<span class="quote">NDB</span>” mean? </p></li><li><p><a href="faqs.html#qandaitem-28-10-2">28.10.2: </a><a class="indexterm" name="id3165530"></a> What's the difference in using Cluster <span class="foreignphrase"><em class="foreignphrase">vs</em></span> using replication? </p></li><li><p><a href="faqs.html#qandaitem-28-10-3">28.10.3: </a><a class="indexterm" name="id3165556"></a> Do I need to do any special networking to run Cluster? How do computers in a cluster communicate? </p></li><li><p><a href="faqs.html#qandaitem-28-10-4">28.10.4: </a><a class="indexterm" name="id3165578"></a> How many computers do I need to run a cluster, and why? </p></li><li><p><a href="faqs.html#qandaitem-28-10-5">28.10.5: </a><a class="indexterm" name="id3165599"></a> What do the different computers do in a MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-6">28.10.6: </a><a class="indexterm" name="id3165620"></a> With which operating systems can I use Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-7">28.10.7: </a><a class="indexterm" name="id3165647"></a> What are the hardware requirements for running MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-8">28.10.8: </a><a class="indexterm" name="id3165669"></a> How much RAM do I need? Is it possible to use disk memory at all? </p></li><li><p><a href="faqs.html#qandaitem-28-10-9">28.10.9: </a> What filesystems can I use with MySQL Cluster? What about network filesystems or network shares? </p></li><li><p><a href="faqs.html#qandaitem-28-10-10">28.10.10: </a><a class="indexterm" name="id3165701"></a> Can I run MySQL Cluster nodes inside virtual machines (such as those created by VMWare, Parallels, or Xen)? </p></li><li><p><a href="faqs.html#qandaitem-28-10-11">28.10.11: </a><a class="indexterm" name="id3165723"></a><a class="indexterm" name="id3165735"></a> I'm trying to populate a Cluster database. The loading process terminates prematurely and I get an error message like this one: </p><pre class="programlisting"><span class="errortext">ERROR 1114: The table 'my_cluster_table' is full</span> </pre><p> Why is this happening? </p></li><li><p><a href="faqs.html#qandaitem-28-10-12">28.10.12: </a><a class="indexterm" name="id3165768"></a> MySQL Cluster uses TCP/IP. Does this mean that I can run it over the Internet, with one or more nodes in remote locations? </p></li><li><p><a href="faqs.html#qandaitem-28-10-13">28.10.13: </a><a class="indexterm" name="id3165791"></a> Do I have to learn a new programming or query language to use MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-14">28.10.14: </a><a class="indexterm" name="id3165813"></a> How do I find out what an error or warning message means when using MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-15">28.10.15: </a><a class="indexterm" name="id3165834"></a> Is MySQL Cluster transaction-safe? What isolation levels are supported? </p></li><li><p><a href="faqs.html#qandaitem-28-10-16">28.10.16: </a> What storage engines are supported by MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-17">28.10.17: </a><a class="indexterm" name="id3165865"></a> Which versions of the MySQL software support Cluster? Do I have to compile from source? </p></li><li><p><a href="faqs.html#qandaitem-28-10-18">28.10.18: </a> In the event of a catastrophic failure — say, for instance, the whole city loses power <span class="emphasis"><em>and</em></span> my UPS fails — would I lose all my data? </p></li><li><p><a href="faqs.html#qandaitem-28-10-19">28.10.19: </a> Is it possible to use <code class="literal">FULLTEXT</code> indexes with Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-20">28.10.20: </a> Can I run multiple nodes on a single computer? </p></li><li><p><a href="faqs.html#qandaitem-28-10-21">28.10.21: </a> Can I add nodes to a cluster without restarting it? </p></li><li><p><a href="faqs.html#qandaitem-28-10-22">28.10.22: </a> Are there any limitations that I should be aware of when using MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-23">28.10.23: </a><a class="indexterm" name="id3165949"></a> How do I import an existing MySQL database into a cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-24">28.10.24: </a> How do cluster nodes communicate with one another? </p></li><li><p><a href="faqs.html#qandaitem-28-10-25">28.10.25: </a><a class="indexterm" name="id3165985"></a><a class="indexterm" name="id3165997"></a> What is an <em class="firstterm">arbitrator</em>? </p></li><li><p><a href="faqs.html#qandaitem-28-10-26">28.10.26: </a><a class="indexterm" name="id3166017"></a> What data types are supported by MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-27">28.10.27: </a><a class="indexterm" name="id3166038"></a> How do I start and stop MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-28">28.10.28: </a> What happens to cluster data when the cluster is shut down? </p></li><li><p><a href="faqs.html#qandaitem-28-10-29">28.10.29: </a> Is it helpful to have more than one management node for a cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-30">28.10.30: </a> Can I mix different kinds of hardware and operating systems in one MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-31">28.10.31: </a> Can I run two data nodes on a single host? Two SQL nodes? </p></li><li><p><a href="faqs.html#qandaitem-28-10-32">28.10.32: </a> Can I use hostnames with MySQL Cluster? </p></li><li><p><a href="faqs.html#qandaitem-28-10-33">28.10.33: </a> How do I handle MySQL users in a Cluster having multiple MySQL servers? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-10-1"></a><span class="bold"><strong>28.10.1: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166128"></a></strong></span><span class="bold"><strong> What does “<span class="quote">NDB</span>” mean? </strong></span></p><p> This stands for “<span class="quote"><span class="bold"><strong>N</strong></span>etwork <span class="bold"><strong>D</strong></span>ata<span class="bold"><strong>b</strong></span>ase</span>”. <code class="literal">NDB</code> (also known as <code class="literal">NDB Cluster</code> or <code class="literal">NDBCLUSTER</code>) is the storage engine that enables clustering in MySQL. </p><p><a name="qandaitem-28-10-2"></a><span class="bold"><strong>28.10.2: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166199"></a></strong></span><span class="bold"><strong> What's the difference in using Cluster <span class="foreignphrase"><em class="foreignphrase">vs</em></span> using replication? </strong></span></p><p> In a replication setup, a master MySQL server updates one or more slaves. Transactions are committed sequentially, and a slow transaction can cause the slave to lag behind the master. This means that if the master fails, it is possible that the slave might not have recorded the last few transactions. If a transaction-safe engine such as <code class="literal">InnoDB</code> is being used, a transaction will either be complete on the slave or not applied at all, but replication does not guarantee that all data on the master and the slave will be consistent at all times. In MySQL Cluster, all data nodes are kept in synchrony, and a transaction committed by any one data node is committed for all data nodes. In the event of a data node failure, all remaining data nodes remain in a consistent state. </p><p> In short, whereas standard MySQL replication is asynchronous, MySQL Cluster is synchronous. </p><p> We have implemented (asynchronous) replication for Cluster in MySQL 5.1. This includes the capability to replicate both between two clusters, and from a MySQL cluster to a non-Cluster MySQL server. However, we do not plan to backport this functionality to MySQL 5.0. </p><p><a name="qandaitem-28-10-3"></a><span class="bold"><strong>28.10.3: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166274"></a></strong></span><span class="bold"><strong> Do I need to do any special networking to run Cluster? How do computers in a cluster communicate? </strong></span></p><p> MySQL Cluster is intended to be used in a high-bandwidth environment, with computers connecting via TCP/IP. Its performance depends directly upon the connection speed between the cluster's computers. The minimum connectivity requirements for Cluster include a typical 100-megabit Ethernet network or the equivalent. We recommend you use gigabit Ethernet whenever available. </p><p> The faster SCI protocol is also supported, but requires special hardware. See <a href="mysql-cluster.html#mysql-cluster-interconnects" title="16.9. Using High-Speed Interconnects with MySQL Cluster">Section 16.9, “Using High-Speed Interconnects with MySQL Cluster”</a>, for more information about SCI. </p><p><a name="qandaitem-28-10-4"></a><span class="bold"><strong>28.10.4: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166327"></a></strong></span><span class="bold"><strong> How many computers do I need to run a cluster, and why? </strong></span></p><p> A minimum of three computers is required to run a viable cluster. However, the minimum <span class="bold"><strong>recommended</strong></span> number of computers in a MySQL Cluster is four: one each to run the management and SQL nodes, and two computers to serve as data nodes. The purpose of the two data nodes is to provide redundancy; the management node must run on a separate machine to guarantee continued arbitration services in the event that one of the data nodes fails. </p><p> To provide increased throughput and high availability, you should use multiple SQL nodes (MySQL Servers connected to the cluster). It is also possible (although not strictly necessary) to run multiple management servers. </p><p><a name="qandaitem-28-10-5"></a><span class="bold"><strong>28.10.5: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166384"></a></strong></span><span class="bold"><strong> What do the different computers do in a MySQL Cluster? </strong></span></p><p> A MySQL Cluster has both a physical and logical organization, with computers being the physical elements. The logical or functional elements of a cluster are referred to as <em class="firstterm">nodes</em>, and a computer housing a cluster node is sometimes referred to as a <em class="firstterm">cluster host</em>. There are three types of nodes, each corresponding to a specific role within the cluster. These are: </p><a class="indexterm" name="id3166420"></a><div class="itemizedlist"><ul type="disc"><li><p> <span class="bold"><strong>Management node (MGM node)</strong></span>: Provides management services for the cluster as a whole, including startup, shutdown, backups, and configuration data for the other nodes. The management node server is implemented as the application <span><strong class="command">ndb_mgmd</strong></span>; the management client used to control MySQL Cluster via the MGM node is <span><strong class="command">ndb_mgm</strong></span>. </p></li><li><p> <span class="bold"><strong>Data node</strong></span>: Stores and replicates data. Data node functionality is handled by an instance of the NDB data node process <span><strong class="command">ndbd</strong></span>. </p></li><li><p> <span class="bold"><strong>SQL node</strong></span>: This is simply an instance of MySQL Server (<span><strong class="command">mysqld</strong></span>) that is built with support for the <code class="literal">NDB Cluster</code> storage engine and started with the <span><strong class="command">--ndb-cluster</strong></span> option to enable the engine. </p></li></ul></div><p><a name="qandaitem-28-10-6"></a><span class="bold"><strong>28.10.6: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166534"></a></strong></span><span class="bold"><strong></strong></span><span class="bold"><strong> With which operating systems can I use Cluster? </strong></span></p><p> MySQL Cluster is supported on most Unix-like operating systems, including Linux, Mac OS X, Solaris, BSD, HP-UX, AIX, and IRIX, among others, as well as Novell Netware. Cluster is not supported for Windows at this time. However, we are working to add Cluster support for other platforms, including Windows, and our goal is to offer MySQL Cluster on all platforms for which MySQL itself is supported. </p><p> For more detailed information concerning the level of support which is offered for MySQL Cluster on various operating system versions, OS distributions, and hardware platforms, please refer to <a href="http://www.mysql.com/support/supportedplatforms/cluster.html" target="_top">http://www.mysql.com/support/supportedplatforms/cluster.html</a>. </p><p><a name="qandaitem-28-10-7"></a><span class="bold"><strong>28.10.7: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166598"></a></strong></span><span class="bold"><strong> What are the hardware requirements for running MySQL Cluster? </strong></span></p><p> Cluster should run on any platform for which NDB-enabled binaries are available. Naturally, faster CPUs and more memory will improve performance, and 64-bit CPUs will likely be more effective than 32-bit processors. There must be sufficient memory on machines used for data nodes to hold each node's share of the database (see <span class="emphasis"><em>How much RAM do I Need?</em></span> for more information). Nodes can communicate via a standard TCP/IP network and hardware. For SCI support, special networking hardware is required. </p><p><a name="qandaitem-28-10-8"></a><span class="bold"><strong>28.10.8: </strong></span><span class="bold"><strong><a class="indexterm" name="id3166652"></a></strong></span><span class="bold"><strong> How much RAM do I need? Is it possible to use disk memory at all? </strong></span></p><p> In MySQL-5.0, Cluster is in-memory only. This means that all table data (including indexes) is stored in RAM. Therefore, if your data takes up 1GB of space and you want to replicate it once in the cluster, you need 2GB of memory to do so (1 GB per replica). This is in addition to the memory required by the operating system and any applications running on the cluster computers. </p><p> If a data node's memory usage exceeds what is available in RAM, then the system will attempt to use swap space up to the limit set for <code class="literal">DataMemory</code>. However, this will at best result in severely degraded performance, and may cause the node to be dropped due to slow response time (missed heartbeats). We do not recommend on relying on disk swapping in a production environment for this reason. In any case, once the <code class="literal">DataMemory</code> limit is reached, any operations requiring additional memory (such as inserts) will fail. </p><p> (We have implemented disk data storage for MySQL Cluster in MySQL 5.1, but we have no plans to add this capability in MySQL 5.0.) </p><p> You can use the following formula for obtaining a rough estimate of how much RAM is needed for each data node in the cluster: </p><pre class="programlisting">(SizeofDatabase × NumberOfReplicas × 1.1 ) / NumberOfDataNodes </pre><p> To calculate the memory requirements more exactly requires determining, for each table in the cluster database, the storage space required per row (see <a href="data-types.html#storage-requirements" title="10.5. Data Type Storage Requirements">Section 10.5, “Data Type Storage Requirements”</a>, for details), and multiplying this by the number of rows. You must also remember to account for any column indexes as follows: </p><div class="itemizedlist"><ul type="disc"><li><p> Each primary key or hash index created for an <code class="literal">NDBCluster</code> table requires 21–25 bytes per record. These indexes use <code class="literal">IndexMemory</code>. </p></li><li><p> Each ordered index requires 10 bytes storage per record, using <code class="literal">DataMemory</code>. </p></li><li><p> Creating a primary key or unique index also creates an ordered index, unless this index is created with <code class="literal">USING HASH</code>. In other words: </p><div class="itemizedlist"><ul type="circle"><li><p> A primary key or unique index on a Cluster table normally takes up 31 to 35 bytes per record. </p></li><li><p> However, if the primary key or unique index is created with <code class="literal">USING HASH</code>, then it requires only 21 to 25 bytes per record. </p></li></ul></div></li></ul></div><p> Note that creating MySQL Cluster tables with <code class="literal">USING HASH</code> for all primary keys and unique indexes will generally cause table updates to run more quickly — in some cases by a much as 20 to 30 percent faster than updates on tables where <code class="literal">USING HASH</code> was not used in creating primary and unique keys. This is due to the fact that less memory is required (because no ordered indexes are created), and that less CPU must be utilized (because fewer indexes must be read and possibly updated). However, it also means that queries that could otherwise use range scans must be satisfied by other means, which can result in slower selects. </p><a class="indexterm" name="id3166858"></a><a class="indexterm" name="id3166871"></a><p> When calculating Cluster memory requirements, you may find useful the <code class="filename">ndb_size.pl</code> utility which is available in recent MySQL 5.0 releases. This Perl script connects to a current (non-Cluster) MySQL database and creates a report on how much space that database would require if it used the <code class="literal">NDBCluster</code> storage engine. For more information, see <a href="mysql-cluster.html#mysql-cluster-utilities-ndb-size" title="16.8.14. ndb_size.pl — NDBCluster Size Requirement Estimator">Section 16.8.14, “<span><strong class="command">ndb_size.pl</strong></span> — NDBCluster Size Requirement Estimator”</a>. </p><p> It is especially important to keep in mind that <span class="emphasis"><em>every MySQL Cluster table must have a primary key</em></span>. The <code class="literal">NDB</code> storage engine creates a primary key automatically if none is defined, and this primary key is created without <code class="literal">USING HASH</code>. </p><p> There is no easy way to determine exactly how much memory is being used for storage of Cluster indexes at any given time; however, warnings are written to the Cluster log when 80% of available <code class="literal">DataMemory</code> or <code class="literal">IndexMemory</code> is in use, and again when use reaches 85%, 90%, and so on. </p><p><a name="qandaitem-28-10-9"></a><span class="bold"><strong>28.10.9: </strong></span><span class="bold"><strong> What filesystems can I use with MySQL Cluster? What about network filesystems or network shares? </strong></span></p><p> Generally, any filesystem that is native to the host operating system should work well with MySQL Cluster. If you find that a given filesystem works particularly well (or not so especially well) with MySQL Cluster, we invite you to discuss your findings in the <a href="http://forums.mysql.com/list.php?25" target="_top">MySQL Cluster Forums</a>. </p><p> We do not test MySQL Cluster with <code class="literal">FAT</code> or <code class="literal">VFAT</code> filesystems on Linux. Because of this, and due to the fact that these are not very useful for any purpose other than sharing disk partitions between Linux and Windows operating systems on multi-boot computers, we do not recommend their use with MySQL Cluster. </p><p> MySQL Cluster is implemented as a shared-nothing solution; the idea behind this is that the failure of a single piece of hardware should not cause the failure of multiple cluster nodes, or possibly even the failure of the cluster as a whole. For this reason, the use of network shares or network filesystems is not supported for MySQL Cluster. This also applies to shared storage devices such as SANs. </p><p><a name="qandaitem-28-10-10"></a><span class="bold"><strong>28.10.10: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167028"></a></strong></span><span class="bold"><strong> Can I run MySQL Cluster nodes inside virtual machines (such as those created by VMWare, Parallels, or Xen)? </strong></span></p><p> This is possible but not recommended for a production environment. </p><p> We have found that running MySQL Cluster processes inside a virtual machine can give rise to issues with timing and disk subsystems that have a strong negative impact on the operation of the cluster. The behavior of the cluster is often unpredictable in these cases. </p><p> If the issue can be reproduced outside the virtual environment, then we may be able to provide assistance. Otherwise, we cannot support it at this time. </p><p><a name="qandaitem-28-10-11"></a><span class="bold"><strong>28.10.11: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167080"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3167096"></a></strong></span><span class="bold"><strong> I'm trying to populate a Cluster database. The loading process terminates prematurely and I get an error message like this one: <pre class="programlisting"><span class="errortext">ERROR 1114: The table 'my_cluster_table' is full</span> </pre> Why is this happening? </strong></span></p><p> The cause is very likely to be that your setup does not provide sufficient RAM for all table data and all indexes, <span class="emphasis"><em>including the primary key required by the <code class="literal">NDB</code> storage engine and automatically created in the event that the table definition does not include the definition of a primary key</em></span>. </p><p> It is also worth noting that all data nodes should have the same amount of RAM, since no data node in a cluster can use more memory than the least amount available to any individual data node. In other words, if there are four computers hosting Cluster data nodes, and three of these have 3GB of RAM available to store Cluster data while the remaining data node has only 1GB RAM, then each data node can devote only 1GB to clustering. </p><p><a name="qandaitem-28-10-12"></a><span class="bold"><strong>28.10.12: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167169"></a></strong></span><span class="bold"><strong> MySQL Cluster uses TCP/IP. Does this mean that I can run it over the Internet, with one or more nodes in remote locations? </strong></span></p><p> It is <span class="emphasis"><em>very</em></span> unlikely that a cluster would perform reliably under such conditions, as MySQL Cluster was designed and implemented with the assumption that it would be run under conditions guaranteeing dedicated high-speed connectivity such as that found in a LAN setting using 100 Mbps or gigabit Ethernet — preferably the latter. We neither test nor warrant its performance using anything slower than this. </p><a class="indexterm" name="id3167212"></a><p> Also, it is extremely important to keep in mind that communications between the nodes in a MySQL Cluster are not secure; they are neither encrypted nor safeguarded by any other protective mechanism. The most secure configuration for a cluster is in a private network behind a firewall, with no direct access to any Cluster data or management nodes from outside. (For SQL nodes, you should take the same precautions as you would with any other instance of the MySQL server.) </p><p><a name="qandaitem-28-10-13"></a><span class="bold"><strong>28.10.13: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167250"></a></strong></span><span class="bold"><strong> Do I have to learn a new programming or query language to use MySQL Cluster? </strong></span></p><p> No. Although some specialized commands are used to manage and configure the cluster itself, only standard (My)SQL queries and commands are required for the following operations: </p><div class="itemizedlist"><ul type="disc"><li><p> Creating, altering, and dropping tables </p></li><li><p> Inserting, updating, and deleting table data </p></li><li><p> Creating, changing, and dropping primary and unique indexes </p></li></ul></div><p> Some specialized configuration parameters and files are required to set up a MySQL Cluster — see <a href="mysql-cluster.html#mysql-cluster-config-file" title="16.3.4. Configuration File">Section 16.3.4, “Configuration File”</a>, for information about these. </p><p> A few simple commands are used in the MySQL Cluster management client for tasks such as starting and stopping cluster nodes. See <a href="mysql-cluster.html#mysql-cluster-mgm-client-commands" title="16.6.2. Commands in the MySQL Cluster Management Client">Section 16.6.2, “Commands in the MySQL Cluster Management Client”</a>. </p><p><a name="qandaitem-28-10-14"></a><span class="bold"><strong>28.10.14: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167347"></a></strong></span><span class="bold"><strong> How do I find out what an error or warning message means when using MySQL Cluster? </strong></span></p><p> There are two ways in which this can be done: </p><div class="itemizedlist"><ul type="disc"><li><a class="indexterm" name="id3167375"></a><a class="indexterm" name="id3167387"></a><p> From within the <span><strong class="command">mysql</strong></span> client, use <span><strong class="command">SHOW ERRORS</strong></span> or <span><strong class="command">SHOW WARNINGS</strong></span> immediately upon being notified of the error or warning condition. Errors and warnings also be displayed in MySQL Query Browser. </p></li><li><a class="indexterm" name="id3167428"></a><p> From a system shell prompt, use <span><strong class="command">perror --ndb <em class="replaceable"><code>error_code</code></em></strong></span>. </p></li></ul></div><p><a name="qandaitem-28-10-15"></a><span class="bold"><strong>28.10.15: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167468"></a></strong></span><span class="bold"><strong> Is MySQL Cluster transaction-safe? What isolation levels are supported? </strong></span></p><p> <span class="emphasis"><em>Yes</em></span>: For tables created with the <code class="literal">NDB</code> storage engine, transactions are supported. In MySQL 5.0, Cluster supports only the <code class="literal">READ COMMITTED</code> transaction isolation level. </p><p><a name="qandaitem-28-10-16"></a><span class="bold"><strong>28.10.16: </strong></span><span class="bold"><strong> What storage engines are supported by MySQL Cluster? </strong></span></p><p> Clustering in MySQL is supported only by the <code class="literal">NDB</code> storage engine. That is, in order for a table to be shared between nodes in a cluster, it must be created using <code class="literal">ENGINE=NDB</code> (or <code class="literal">ENGINE=NDBCLUSTER</code>, which is equivalent). </p><p> It is possible to create tables using other storage engines (such as <code class="literal">MyISAM</code> or <code class="literal">InnoDB</code>) on a MySQL server being used for clustering, but these non-<code class="literal">NDB</code> tables will <span class="bold"><strong>not</strong></span> participate in the cluster; they are local to the individual MySQL server instance on which they are created. </p><p><a name="qandaitem-28-10-17"></a><span class="bold"><strong>28.10.17: </strong></span><span class="bold"><strong><a class="indexterm" name="id3167592"></a></strong></span><span class="bold"><strong> Which versions of the MySQL software support Cluster? Do I have to compile from source? </strong></span></p><p> Cluster is supported in all server binaries in the 5.0 release series for operating systems on which MySQL Cluster is available. See <a href="server-administration.html#mysqld" title="5.2. mysqld — The MySQL Server">Section 5.2, “<span><strong class="command">mysqld</strong></span> — The MySQL Server”</a>. You can determine whether your server has NDB support using either the <code class="literal">SHOW VARIABLES LIKE 'have_%'</code> or <code class="literal">SHOW ENGINES</code> statement. </p><p> Linux users, please note that <code class="literal">NDB</code> is <span class="emphasis"><em>not</em></span> included in the standard MySQL server RPMs. Beginning with MySQL 5.0.4, there are separate RPM packages for the NDB storage engine and accompanying management and other tools; see the NDB RPM Downloads section of the MySQL 5.0 Downloads page for these. (Prior to 5.0.4, you had to use the <code class="literal">-max</code> binaries supplied as <code class="filename">.tar.gz</code> archives. This is still possible, but is not required, so you can use your Linux distribution's RPM manager if you prefer.) </p><p> You can also obtain NDB support by compiling MySQL from source, but it is not necessary to do so simply to use MySQL Cluster. To download the latest binary, RPM, or source distribution in the MySQL 5.0 series, visit <a href="http://dev.mysql.com/downloads/mysql/5.0.html" target="_top">http://dev.mysql.com/downloads/mysql/5.0.html</a>. </p><p><a name="qandaitem-28-10-18"></a><span class="bold"><strong>28.10.18: </strong></span><span class="bold"><strong> In the event of a catastrophic failure — say, for instance, the whole city loses power <span class="emphasis"><em>and</em></span> my UPS fails — would I lose all my data? </strong></span></p><p> All committed transactions are logged. Therefore, although it is possible that some data could be lost in the event of a catastrophe, this should be quite limited. Data loss can be further reduced by minimizing the number of operations per transaction. (It is not a good idea to perform large numbers of operations per transaction in any case.) </p><p><a name="qandaitem-28-10-19"></a><span class="bold"><strong>28.10.19: </strong></span><span class="bold"><strong> Is it possible to use <code class="literal">FULLTEXT</code> indexes with Cluster? </strong></span></p><p> <code class="literal">FULLTEXT</code> indexing is not supported by the <code class="literal">NDB</code> storage engine in MySQL 5.0, or by any storage engine other than <code class="literal">MyISAM</code>. We are working to add this capability in a future release. </p><p><a name="qandaitem-28-10-20"></a><span class="bold"><strong>28.10.20: </strong></span><span class="bold"><strong> Can I run multiple nodes on a single computer? </strong></span></p><p> It is possible but not advisable. One of the chief reasons to run a cluster is to provide redundancy. To enjoy the full benefits of this redundancy, each node should reside on a separate machine. If you place multiple nodes on a single machine and that machine fails, you lose all of those nodes. Given that MySQL Cluster can be run on commodity hardware loaded with a low-cost (or even no-cost) operating system, the expense of an extra machine or two is well worth it to safeguard mission-critical data. It also worth noting that the requirements for a cluster host running a management node are minimal. This task can be accomplished with a 200 MHz Pentium CPU and sufficient RAM for the operating system plus a small amount of overhead for the <span><strong class="command">ndb_mgmd</strong></span> and <span><strong class="command">ndb_mgm</strong></span> processes. </p><p> It is acceptable to run multiple cluster data nodes on a single host for learning about MySQL Cluster, or for testing purposes; however, this is <span class="emphasis"><em>not</em></span> supported for production use. </p><p><a name="qandaitem-28-10-21"></a><span class="bold"><strong>28.10.21: </strong></span><span class="bold"><strong> Can I add nodes to a cluster without restarting it? </strong></span></p><p> Not at present. A simple restart is all that is required for adding new MGM or SQL nodes to a Cluster. When adding data nodes the process is more complex, and requires the following steps: </p><div class="orderedlist"><ol type="1"><li><p> Make a complete backup of all Cluster data. </p></li><li><p> Completely shut down the cluster and all cluster node processes. </p></li><li><p> Restart the cluster, using the <code class="option">--initial</code> startup option. </p><div class="warning" style="margin-left: 0.5in; margin-right: 0.5in;"><h3 class="title">Warning</h3><p> Never use the <code class="option">--initial</code> when starting <span><strong class="command">ndbd</strong></span> except when necessary to clear the data node filesystem. See <a href="mysql-cluster.html#mysql-cluster-ndbd-command-options" title="16.5.5.2. Command Options for ndbd">Section 16.5.5.2, “Command Options for <span><strong class="command">ndbd</strong></span>”</a>, for information about when this is required. </p></div></li><li><p> Restore all cluster data from the backup. </p></li></ol></div><p> In a future MySQL Cluster release series, we hope to implement a “<span class="quote">hot</span>” reconfiguration capability for MySQL Cluster to minimize (if not eliminate) the requirement for restarting the cluster when adding new nodes. However, this is not planned for MySQL 5.0. </p><p><a name="qandaitem-28-10-22"></a><span class="bold"><strong>28.10.22: </strong></span><span class="bold"><strong> Are there any limitations that I should be aware of when using MySQL Cluster? </strong></span></p><p> Limitations on <code class="literal">NDB</code> tables in MySQL 5.0 include: </p><div class="itemizedlist"><ul type="disc"><li><p> Temporary tables are not supported; a <code class="literal">CREATE TEMPORARY TABLE</code> statement using <code class="literal">ENGINE=NDB</code> or <code class="literal">ENGINE=NDBCLUSTER</code> fails with an error. </p></li><li><p> <code class="literal">FULLTEXT</code> indexes and index prefixes are not supported. Only complete columns may be indexed. </p></li><li><p> Spatial data types are not supported. See <a href="spatial-extensions.html" title="Chapter 17. Spatial Extensions">Chapter 17, <i>Spatial Extensions</i></a>. </p></li><li><p> Only complete rollbacks for transactions are supported. Partial rollbacks and rollbacks to save points are not supported. </p></li><li><p> The maximum number of attributes allowed per table is 128, and attribute names cannot be any longer than 31 characters. For each table, the maximum combined length of the table and database names is 122 characters. </p></li><li><p> The maximum size for a table row is 8 kilobytes, not counting <code class="literal">BLOB</code> values. There is no set limit for the number of rows per table. Table size limits depend on a number of factors, in particular on the amount of RAM available to each data node. </p></li><li><p> The <code class="literal">NDB</code> engine does not support foreign key constraints. As with <code class="literal">MyISAM</code> tables, these are ignored. </p></li></ul></div><p> For a complete listing of limitations in MySQL Cluster, see <a href="mysql-cluster.html#mysql-cluster-limitations" title="16.10. Known Limitations of MySQL Cluster">Section 16.10, “Known Limitations of MySQL Cluster”</a>. </p><p><a name="qandaitem-28-10-23"></a><span class="bold"><strong>28.10.23: </strong></span><span class="bold"><strong><a class="indexterm" name="id3168088"></a></strong></span><span class="bold"><strong> How do I import an existing MySQL database into a cluster? </strong></span></p><p> You can import databases into MySQL Cluster much as you would with any other version of MySQL. Other than the limitations mentioned elsewhere in this FAQ and in <a href="mysql-cluster.html#mysql-cluster-limitations" title="16.10. Known Limitations of MySQL Cluster">Section 16.10, “Known Limitations of MySQL Cluster”</a>, the only other special requirement is that any tables to be included in the cluster must use the <code class="literal">NDB</code> storage engine. This means that the tables must be created with <code class="literal">ENGINE=NDB</code> or <code class="literal">ENGINE=NDBCLUSTER</code>. </p><p> It is also possible to convert existing tables using other storage engines to <code class="literal">NDB Cluster</code> using one or more <code class="literal">ALTER TABLE</code> statement, but this requires an additional workaround. See <a href="mysql-cluster.html#mysql-cluster-limitations" title="16.10. Known Limitations of MySQL Cluster">Section 16.10, “Known Limitations of MySQL Cluster”</a>, for details. </p><p><a name="qandaitem-28-10-24"></a><span class="bold"><strong>28.10.24: </strong></span><span class="bold"><strong></strong></span><span class="bold"><strong> How do cluster nodes communicate with one another? </strong></span></p><p> Cluster nodes can communicate via any of three different protocols: TCP/IP, SHM (shared memory), and SCI (Scalable Coherent Interface). Where available, SHM is used by default between nodes residing on the same cluster host; however, this is considered experimental in MySQL 5.0. SCI is a high-speed (1 gigabit per second and higher), high-availability protocol used in building scalable multi-processor systems; it requires special hardware and drivers. See <a href="mysql-cluster.html#mysql-cluster-interconnects" title="16.9. Using High-Speed Interconnects with MySQL Cluster">Section 16.9, “Using High-Speed Interconnects with MySQL Cluster”</a>, for more about using SCI as a transport mechanism in MySQL Cluster. </p><p><a name="qandaitem-28-10-25"></a><span class="bold"><strong>28.10.25: </strong></span><span class="bold"><strong><a class="indexterm" name="id3168218"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3168233"></a></strong></span><span class="bold"><strong> What is an <em class="firstterm">arbitrator</em>? </strong></span></p><p> If one or more nodes in a cluster fail, it is possible that not all cluster nodes will be able to “<span class="quote">see</span>” one another. In fact, it is possible that two sets of nodes might become isolated from one another in a network partitioning, also known as a “<span class="quote">split brain</span>” scenario. This type of situation is undesirable because each set of nodes tries to behave as though it is the entire cluster. </p><p> When cluster nodes go down, there are two possibilities. If more than 50% of the remaining nodes can communicate with each other, we have what is sometimes called a “<span class="quote">majority rules</span>” situation, and this set of nodes is considered to be the cluster. The arbitrator comes into play when there is an even number of nodes: in such cases, the set of nodes to which the arbitrator belongs is considered to be the cluster, and nodes not belonging to this set are shut down. </p><p> The preceding information is somewhat simplified. A more complete explanation taking into account node groups follows: </p><p> When all nodes in at least one node group are alive, network partitioning is not an issue, because no one portion of the cluster can form a functional cluster. The real problem arises when no single node group has all its nodes alive, in which case network partitioning (the “<span class="quote">split-brain</span>” scenario) becomes possible. Then an arbitrator is required. All cluster nodes recognize the same node as the arbitrator, which is normally the management server; however, it is possible to configure any of the MySQL Servers in the cluster to act as the arbitrator instead. The arbitrator accepts the first set of cluster nodes to contact it, and tells the remaining set to shut down. Arbitrator selection is controlled by the <code class="literal">ArbitrationRank</code> configuration parameter for MySQL Server and management server nodes. (See <a href="mysql-cluster.html#mysql-cluster-mgm-definition" title="16.3.4.4. Defining the Management Server">Section 16.3.4.4, “Defining the Management Server”</a>, for details.) It should also be noted that the role of arbitrator does not in and of itself impose any heavy demands upon the host so designated, and thus the arbitrator host does not need to be particularly fast or to have extra memory especially for this purpose. </p><p><a name="qandaitem-28-10-26"></a><span class="bold"><strong>28.10.26: </strong></span><span class="bold"><strong><a class="indexterm" name="id3168343"></a></strong></span><span class="bold"><strong> What data types are supported by MySQL Cluster? </strong></span></p><p> MySQL Cluster supports all of the usual MySQL data types, with the exception of those associated with MySQL's spatial extensions. (See <a href="spatial-extensions.html" title="Chapter 17. Spatial Extensions">Chapter 17, <i>Spatial Extensions</i></a>.) In addition, there are some differences with regard to indexes when used with <code class="literal">NDB</code> tables. </p><div class="note" style="margin-left: 0.5in; margin-right: 0.5in;"><h3 class="title">Note</h3><p> MySQL Cluster tables (that is, tables created with <code class="literal">ENGINE=NDBCLUSTER</code>) have only fixed-width rows. This means that (for example) each record containing a <code class="literal">VARCHAR(255)</code> column will require space for 255 characters (as required for the character set and collation being used for the table), regardless of the actual number of characters stored therein. This issue is expected to be fixed in a future MySQL release series. </p></div><p> See <a href="mysql-cluster.html#mysql-cluster-limitations" title="16.10. Known Limitations of MySQL Cluster">Section 16.10, “Known Limitations of MySQL Cluster”</a>, for more information about these issues. </p><p><a name="qandaitem-28-10-27"></a><span class="bold"><strong>28.10.27: </strong></span><span class="bold"><strong><a class="indexterm" name="id3168428"></a></strong></span><span class="bold"><strong> How do I start and stop MySQL Cluster? </strong></span></p><p> It is necessary to start each node in the cluster separately, in the following order: </p><div class="orderedlist"><ol type="1"><li><p> Start the management node with the <span><strong class="command">ndb_mgmd</strong></span> command. </p></li><li><p> Start each data node with the <span><strong class="command">ndbd</strong></span> command. </p></li><li><p> Start each MySQL server (SQL node) using <span><strong class="command">mysqld_safe --user=mysql &</strong></span>. </p></li></ol></div><p> Each of these commands must be run from a system shell on the machine housing the affected node. (You do not have to be physically present at the machine — a remote login shell can be used for this purpose.) You can verify that the cluster is running by starting the MGM management client <span><strong class="command">ndb_mgm</strong></span> on the machine housing the MGM node and issuing the <code class="literal">SHOW</code> or <code class="literal">ALL STATUS</code> command. </p><p> To shut down a running cluster, issue the command <code class="literal">SHUTDOWN</code> in the MGM client. Alternatively, you may enter the following command in a system shell on the machine hosting the MGM node: </p><pre class="programlisting">shell> <strong class="userinput"><code>ndb_mgm -e "SHUTDOWN"</code></strong> </pre><p> (Note that the quotation marks are optional here; the <code class="literal">SHUTDOWN</code> command itself is not case-sensitive.) </p><p> Either of these commands causes the <span><strong class="command">ndb_mgm</strong></span>, <span><strong class="command">ndb_mgm</strong></span>, and any <span><strong class="command">ndbd</strong></span> processes to terminate gracefully. MySQL servers running as Cluster SQL nodes can be stopped using <span><strong class="command">mysqladmin shutdown</strong></span>. </p><p> For more information, see <a href="mysql-cluster.html#mysql-cluster-mgm-client-commands" title="16.6.2. Commands in the MySQL Cluster Management Client">Section 16.6.2, “Commands in the MySQL Cluster Management Client”</a>, and <a href="mysql-cluster.html#mysql-cluster-multi-shutdown-restart" title="16.2.6. Safe Shutdown and Restart">Section 16.2.6, “Safe Shutdown and Restart”</a>. </p><p><a name="qandaitem-28-10-28"></a><span class="bold"><strong>28.10.28: </strong></span><span class="bold"><strong> What happens to cluster data when the cluster is shut down? </strong></span></p><p> The data that was held in memory by the cluster's data nodes is written to disk, and is reloaded into memory the next time that the cluster is started. </p><p><a name="qandaitem-28-10-29"></a><span class="bold"><strong>28.10.29: </strong></span><span class="bold"><strong> Is it helpful to have more than one management node for a cluster? </strong></span></p><p> It can be helpful as a fail-safe. Only one MGM node controls the cluster at any given time, but it is possible to configure one MGM as primary, and one or more additional management nodes to take over in the event that the primary MGM node fails. </p><p> See <a href="mysql-cluster.html#mysql-cluster-config-file" title="16.3.4. Configuration File">Section 16.3.4, “Configuration File”</a>, for information on how to configure MySQL Cluster management nodes. </p><p><a name="qandaitem-28-10-30"></a><span class="bold"><strong>28.10.30: </strong></span><span class="bold"><strong> Can I mix different kinds of hardware and operating systems in one MySQL Cluster? </strong></span></p><p> Yes, so long as all machines and operating systems have the same “<span class="quote">endianness</span>” (all big-endian or all little-endian). It is also possible to use different MySQL Cluster releases on different nodes. However, we recommend this be done only as part of a rolling upgrade procedure (see <a href="mysql-cluster.html#mysql-cluster-rolling-restart" title="16.4.1. Performing a Rolling Restart of the Cluster">Section 16.4.1, “Performing a Rolling Restart of the Cluster”</a>). </p><p><a name="qandaitem-28-10-31"></a><span class="bold"><strong>28.10.31: </strong></span><span class="bold"><strong> Can I run two data nodes on a single host? Two SQL nodes? </strong></span></p><p> Yes, it is possible to do this. In the case of multiple data nodes, it is advisable (but not required) for each node to use a different data directory. If you want to run multiple SQL nodes on one machine, each instance of <span><strong class="command">mysqld</strong></span> must use a different TCP/IP port. However, <span class="emphasis"><em>running more than one cluster node of a given type per machine is not supported for production use</em></span>. </p><p><a name="qandaitem-28-10-32"></a><span class="bold"><strong>28.10.32: </strong></span><span class="bold"><strong> Can I use hostnames with MySQL Cluster? </strong></span></p><p> Yes, it is possible to use DNS and DHCP for cluster hosts. However, if your application requires “<span class="quote">five nines</span>” availability, we recommend using fixed IP addresses. Making communication between Cluster hosts dependent on services such as DNS and DHCP introduces additional points of failure, and the fewer of these, the better. </p><p><a name="qandaitem-28-10-33"></a><span class="bold"><strong>28.10.33: </strong></span><span class="bold"><strong> How do I handle MySQL users in a Cluster having multiple MySQL servers? </strong></span></p><p> MySQL user accounts and privileges are not automatically propagated between different MySQL servers accessing the same MySQL Cluster. Therefore, you must make sure that these are copied between the SQL nodes yourself. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-cjk"></a>A.11. MySQL 5.0 FAQ — MySQL Chinese, Japanese, and Korean Character Sets</h2></div></div></div><a class="indexterm" name="id3168808"></a><a class="indexterm" name="id3168820"></a><a class="indexterm" name="id3168833"></a><a class="indexterm" name="id3168846"></a><p> This set of Frequently Asked Questions derives from the experience of MySQL's Support and Development groups in handling many inquiries about CJK (Chinese-Japanese-Korean) issues. </p><a class="indexterm" name="id3168866"></a><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-11-1">28.11.1: </a> I have inserted CJK characters into my table. Why does <code class="literal">SELECT</code> display them as “<span class="quote">?</span>” characters? </p></li><li><p><a href="faqs.html#qandaitem-28-11-2">28.11.2: </a><a class="indexterm" name="id3168913"></a><a class="indexterm" name="id3168925"></a><a class="indexterm" name="id3168937"></a> What GB (Chinese) character sets does MySQL support? </p></li><li><p><a href="faqs.html#qandaitem-28-11-3">28.11.3: </a><a class="indexterm" name="id3168969"></a><a class="indexterm" name="id3168981"></a><a class="indexterm" name="id3168993"></a> What problems should I be aware of when working with the Big5 Chinese character set? </p></li><li><p><a href="faqs.html#qandaitem-28-11-4">28.11.4: </a><a class="indexterm" name="id3169011"></a><a class="indexterm" name="id3169023"></a><a class="indexterm" name="id3169035"></a> Why do Japanese character set conversions fail? </p></li><li><p><a href="faqs.html#qandaitem-28-11-5">28.11.5: </a> What should I do if I want to convert SJIS <code class="literal">81CA</code> to <code class="literal">cp932</code>? </p></li><li><p><a href="faqs.html#qandaitem-28-11-6">28.11.6: </a><a class="indexterm" name="id3169078"></a><a class="indexterm" name="id3169090"></a><a class="indexterm" name="id3169102"></a> How does MySQL represent the Yen (<code class="literal">¥</code>) sign? </p></li><li><p><a href="faqs.html#qandaitem-28-11-7">28.11.7: </a> Do MySQL plan to make a separate character set where <code class="literal">5C</code> is the Yen sign, as at least one other major DBMS does? </p></li><li><p><a href="faqs.html#qandaitem-28-11-8">28.11.8: </a><a class="indexterm" name="id3169141"></a><a class="indexterm" name="id3169154"></a><a class="indexterm" name="id3169166"></a> Of what issues should I be aware when working with Korean character sets in MySQL? </p></li><li><p><a href="faqs.html#qandaitem-28-11-9">28.11.9: </a><a class="indexterm" name="id3169184"></a><a class="indexterm" name="id3169196"></a><a class="indexterm" name="id3169208"></a> Why do I get <span class="errortext">Data truncated</span> error messages? </p></li><li><p><a href="faqs.html#qandaitem-28-11-10">28.11.10: </a><a class="indexterm" name="id3169229"></a><a class="indexterm" name="id3169241"></a> Why does my GUI front end or browser not display CJK characters correctly in my application using Access, PHP, or another API? </p></li><li><p><a href="faqs.html#qandaitem-28-11-11">28.11.11: </a><a class="indexterm" name="id3169264"></a> I've upgraded to MySQL 5.0. How can I revert to behavior like that in MySQL 4.0 with regard to character sets? </p></li><li><p><a href="faqs.html#qandaitem-28-11-12">28.11.12: </a><a class="indexterm" name="id3169287"></a><a class="indexterm" name="id3169299"></a> Why do some <code class="literal">LIKE</code> and <code class="literal">FULLTEXT</code> searches with CJK characters fail? </p></li><li><p><a href="faqs.html#qandaitem-28-11-13">28.11.13: </a><a class="indexterm" name="id3169332"></a><a class="indexterm" name="id3169344"></a> What CJK character sets are available in MySQL? </p></li><li><p><a href="faqs.html#qandaitem-28-11-14">28.11.14: </a><a class="indexterm" name="id3169365"></a><a class="indexterm" name="id3169378"></a> How do I know whether character <em class="replaceable"><code>X</code></em> is available in all character sets? </p></li><li><p><a href="faqs.html#qandaitem-28-11-15">28.11.15: </a><a class="indexterm" name="id3169403"></a><a class="indexterm" name="id3169415"></a><a class="indexterm" name="id3169427"></a><a class="indexterm" name="id3169438"></a> Why don't CJK strings sort correctly in Unicode? (I) </p></li><li><p><a href="faqs.html#qandaitem-28-11-16">28.11.16: </a><a class="indexterm" name="id3169460"></a><a class="indexterm" name="id3169472"></a><a class="indexterm" name="id3169483"></a><a class="indexterm" name="id3169495"></a><a class="indexterm" name="id3169507"></a> Why don't CJK strings sort correctly in Unicode? (II) </p></li><li><p><a href="faqs.html#qandaitem-28-11-17">28.11.17: </a><a class="indexterm" name="id3169528"></a> Why are my supplementary characters rejected by MySQL? </p></li><li><p><a href="faqs.html#qandaitem-28-11-18">28.11.18: </a><a class="indexterm" name="id3169550"></a><a class="indexterm" name="id3169561"></a><a class="indexterm" name="id3169573"></a> Shouldn't it be “<span class="quote">CJKV</span>”? </p></li><li><p><a href="faqs.html#qandaitem-28-11-19">28.11.19: </a><a class="indexterm" name="id3169594"></a> Does MySQL allow CJK characters to be used in database and table names? </p></li><li><p><a href="faqs.html#qandaitem-28-11-20">28.11.20: </a><a class="indexterm" name="id3169615"></a><a class="indexterm" name="id3169627"></a><a class="indexterm" name="id3169639"></a><a class="indexterm" name="id3169651"></a><a class="indexterm" name="id3169663"></a><a class="indexterm" name="id3169674"></a> Where can I find translations of the MySQL Manual into Chinese, Japanese, and Korean? </p></li><li><p><a href="faqs.html#qandaitem-28-11-21">28.11.21: </a> Where can I get help with CJK and related issues in MySQL? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-11-1"></a><span class="bold"><strong>28.11.1: </strong></span><span class="bold"><strong> I have inserted CJK characters into my table. Why does <code class="literal">SELECT</code> display them as “<span class="quote">?</span>” characters? </strong></span></p><p> This problem is usually due to a setting in MySQL that doesn't match the settings for the application program or the operating system. Here are some common steps for correcting these types of issues: </p><div class="itemizedlist"><ul type="disc"><li><p> <span class="emphasis"><em>Be certain of what MySQL version you are using</em></span>. </p><p> Use the statement <code class="literal">SELECT VERSION();</code> to determine this. </p></li><li><p> <span class="emphasis"><em>Make sure that the database is actually using the desired character set</em></span>. </p><p> People often think that the client character set is always the same as either the server character set or the character set used for display purposes. However, both of these are false assumptions. You can make sure by checking the result of <code class="literal">SHOW CREATE TABLE <em class="replaceable"><code>tablename</code></em></code> or — better — yet by using this statement: </p><pre class="programlisting">SELECT character_set_name, collation_name FROM information_schema.columns WHERE table_schema = your_database_name AND table_name = your_table_name AND column_name = your_column_name; </pre><p> </p></li><li><p> <span class="emphasis"><em>Determine the hexadecimal value of the character or characters that are not being displayed correctly</em></span>. </p><p> You can obtain this information for a column <em class="replaceable"><code>column_name</code></em> in the table <em class="replaceable"><code>table_name</code></em> using the following query: </p><pre class="programlisting">SELECT HEX(<em class="replaceable"><code>column_name</code></em>) FROM <em class="replaceable"><code>table_name</code></em>; </pre><p> <code class="literal">3F</code> is the encoding for the <code class="literal">?</code> character; this means that <code class="literal">?</code> is the character actually stored in the column. This most often happens because of a problem converting a particular character from your client character set to the target character set. </p></li><li><p> <span class="emphasis"><em>Make sure that a round trip possible — that is, when you select <em class="replaceable"><code>literal</code></em> (or <em class="replaceable"><code>_introducer hexadecimal-value</code></em>), you obtain <em class="replaceable"><code>literal</code></em> as a result</em></span>. </p><p> For example, the Japanese <span class="foreignphrase"><em class="foreignphrase">Katakana</em></span> character <span class="foreignphrase"><em class="foreignphrase">Pe</em></span> (<code class="literal">ペ'</code>) exists in all CJK character sets, and has the code point value (hexadecimal coding) <code class="literal">0x30da</code>. To test a round trip for this character, use this query: </p><pre class="programlisting">SELECT 'ペ' AS `ペ`; /* or SELECT _ucs2 0x30da; */ </pre><p> If the result is not also <code class="literal">ペ</code>, then the round trip has failed. </p><p> For bug reports regarding such failures, we might ask you to follow up with <code class="literal">SELECT HEX('ペ');</code>. Then we can determine whether the client encoding is correct. </p></li><li><p> <span class="emphasis"><em>Make sure that the problem is not with the browser or other application, rather than with MySQL</em></span>. </p><p> Use the <span><strong class="command">mysql</strong></span> client program (on Windows: <span><strong class="command">mysql.exe</strong></span>) to accomplish this task. If <span><strong class="command">mysql</strong></span> displays correctly but your application doesn't, then your problem is probably due to system settings. </p><p> To find out what your settings are, use the <code class="literal">SHOW VARIABLES</code> statement, whose output should resemble what is shown here: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SHOW VARIABLES LIKE 'char%';</code></strong> +--------------------------+----------------------------------------+ | Variable_name | Value | +--------------------------+----------------------------------------+ | character_set_client | utf8 | | character_set_connection | utf8 | | character_set_database | latin1 | | character_set_filesystem | binary | | character_set_results | utf8 | | character_set_server | latin1 | | character_set_system | utf8 | | character_sets_dir | /usr/local/mysql/share/mysql/charsets/ | +--------------------------+----------------------------------------+ 8 rows in set (0.03 sec) </pre><p> These are typical character-set settings for an international-oriented client (notice the use of <code class="literal">utf8</code> Unicode) connected to a server in the West (<code class="literal">latin1</code> is a West Europe character set and a default for MySQL). </p><p> Although Unicode (usually the <code class="literal">utf8</code> variant on Unix, and the <code class="literal">ucs2</code> variant on Windows) is preferable to Latin, it's often not what your operating system utilities support best. Many Windows users find that a Microsoft character set, such as <code class="literal">cp932</code> for Japanese Windows, is what's suitable. </p><p> If you cannot control the server settings, and you have no idea what your underlying computer is, then try changing to a common character set for the country that you're in (<code class="literal">euckr</code> = Korea; <code class="literal">gb2312</code> or <code class="literal">gbk</code> = People's Republic of China; <code class="literal">big5</code> = Taiwan; <code class="literal">sjis</code>, <code class="literal">ujis</code>, <code class="literal">cp932</code>, or <code class="literal">eucjpms</code> = Japan; <code class="literal">ucs2</code> or <code class="literal">utf8</code> = anywhere). Usually it is necessary to change only the client and connection and results settings. There is a simple statement which changes all three at once: <code class="literal">SET NAMES</code>. For example: </p><pre class="programlisting">SET NAMES 'big5'; </pre><p> Once the setting is correct, you can make it permanent by editing <code class="filename">my.cnf</code> or <code class="filename">my.ini</code>. For example you might add lines looking like these: </p><pre class="programlisting">[mysqld] character-set-server=big5 [client] default-character-set=big5 </pre><p> </p><p> It is also possible that there are issues with the API configuration setting being used in your application; see <em class="citetitle">Why does my GUI front end or browser not display CJK characters correctly...?</em> for more information. </p></li></ul></div><p><a name="qandaitem-28-11-2"></a><span class="bold"><strong>28.11.2: </strong></span><span class="bold"><strong><a class="indexterm" name="id3170210"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3170225"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3170240"></a></strong></span><span class="bold"><strong></strong></span><span class="bold"><strong></strong></span><span class="bold"><strong> What GB (Chinese) character sets does MySQL support? </strong></span></p><p> MySQL supports the two common variants of the <em class="firstterm">GB</em> (<span class="foreignphrase"><em class="foreignphrase">Guojia Biaozhun</em></span>, or <span class="emphasis"><em>National Standard</em></span>) character sets which are official in the People's Republic of China: <code class="literal">gb2312</code> and <code class="literal">gbk</code>. Sometimes people try to insert <code class="literal">gbk</code> characters into <code class="literal">gb2312</code>, and it works most of the time because <code class="literal">gbk</code> is a superset of <code class="literal">gb2312</code> — but eventually they try to insert a rarer Chinese character and it doesn't work. (See <a href="http://bugs.mysql.com/16072" target="_top">Bug#16072</a> for an example). </p><p> Here, we try to clarify exactly what characters are legitimate in <code class="literal">gb2312</code> or <code class="literal">gbk</code>, with reference to the official documents. Please check these references before reporting <code class="literal">gb2312</code> or <code class="literal">gbk</code> bugs. </p><div class="itemizedlist"><ul type="disc"><li><p> For a complete listing of the <code class="literal">gb2312</code> characters, ordered according to the <code class="literal">gb2312_chinese_ci</code> collation: <a href="http://d.udm.net/bar/~bar/charts/gb2312_chinese_ci.html" target="_top">http://d.udm.net/bar/~bar/charts/gb2312_chinese_ci.html</a>. </p></li><li><p> MySQL's <code class="literal">gbk</code> is in reality “<span class="quote">Microsoft code page 936</span>”. This differs from the official <code class="literal">gbk</code> for characters <code class="literal">A1A4</code> (middle dot), <code class="literal">A1AA</code> (em dash), <code class="literal">A6E0-A6F5</code>, and <code class="literal">A8BB-A8C0</code>. For a listing of the differences, see <a href="http://recode.progiciels-bpi.ca/showfile.html?name=dist/libiconv/gbk.h" target="_top">http://recode.progiciels-bpi.ca/showfile.html?name=dist/libiconv/gbk.h</a>. </p></li><li><p> For a listing of <code class="literal">gbk</code>/Unicode mappings, see <a href="http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP936.TXT" target="_top">http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP936.TXT</a>. </p></li><li><p> For MySQL's listing of <code class="literal">gbk</code> characters, see <a href="http://d.udm.net/bar/~bar/charts/gbk_chinese_ci.html" target="_top">http://d.udm.net/bar/~bar/charts/gbk_chinese_ci.html</a>. </p></li></ul></div><p> </p><p><a name="qandaitem-28-11-3"></a><span class="bold"><strong>28.11.3: </strong></span><span class="bold"><strong><a class="indexterm" name="id3170509"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3170524"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3170539"></a></strong></span><span class="bold"><strong> What problems should I be aware of when working with the Big5 Chinese character set? </strong></span></p><p> MySQL supports the Big5 character set which is common in Hong Kong and Taiwan (Republic of China). MySQL's <code class="literal">big5</code> is in reality Microsoft code page 950, which is very similar to the original <code class="literal">big5</code> character set. We changed to this character set starting with MySQL version 4.1.16 / 5.0.16 (as a result of <a href="http://bugs.mysql.com/12476" target="_top">Bug#12476</a>). For example, the following statements work in current versions of MySQL, but not in old versions: </p><pre class="programlisting">mysql> <strong class="userinput"><code>CREATE TABLE big5 (BIG5 CHAR(1) CHARACTER SET BIG5);</code></strong> Query OK, 0 rows affected (0.13 sec) mysql> <strong class="userinput"><code>INSERT INTO big5 VALUES (0xf9dc);</code></strong> Query OK, 1 row affected (0.00 sec) mysql> <strong class="userinput"><code>SELECT * FROM big5;</code></strong> +------+ | big5 | +------+ | 嫺 | +------+ 1 row in set (0.02 sec) </pre><p> A feature request for adding <code class="literal">HKSCS</code> extensions has been filed. People who need this extension may find the suggested patch for <a href="http://bugs.mysql.com/13577" target="_top">Bug#13577</a> to be of interest. </p><p><a name="qandaitem-28-11-4"></a><span class="bold"><strong>28.11.4: </strong></span><span class="bold"><strong><a class="indexterm" name="id3170649"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3170664"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3170680"></a></strong></span><span class="bold"><strong> Why do Japanese character set conversions fail? </strong></span></p><p> MySQL supports the <code class="literal">sjis</code>, <code class="literal">ujis</code>, <code class="literal">cp932</code>, and <code class="literal">eucjpms</code> character sets, as well as Unicode. A common need is to convert between character sets. For example, there might be a Unix server (typically with <code class="literal">sjis</code> or <code class="literal">ujis</code>) and a Windows client (typically with <code class="literal">cp932</code>). </p><p> In the following conversion table, the <code class="literal">ucs2</code> column represents the source, and the <code class="literal">sjis</code>, <code class="literal">cp932</code>, <code class="literal">ujis</code>, and <code class="literal">eucjpms</code> columns represent the destinations — that is, the last 4 columns provide the hexadecimal result when we use <code class="literal">CONVERT(ucs2)</code> or we assign a <code class="literal">ucs2</code> column containing the value to an <code class="literal">sjis</code>, <code class="literal">cp932</code>, <code class="literal">ujis</code>, or <code class="literal">eucjpms</code> column. </p><div class="informaltable"><table border="1"><colgroup><col><col align="left"><col align="left"><col align="left"><col align="left"><col align="left"></colgroup><thead><tr><th>Character Name</th><th align="left">ucs2</th><th align="left">sjis</th><th align="left">cp932</th><th align="left">ujis</th><th align="left">eucjpms</th></tr></thead><tbody><tr><td>BROKEN BAR</td><td align="left">00A6</td><td align="left">3F</td><td align="left">3F</td><td align="left">8FA2C3</td><td align="left">3F</td></tr><tr><td>FULLWIDTH BROKEN BAR</td><td align="left">FFE4</td><td align="left">3F</td><td align="left">FA55</td><td align="left">3F</td><td align="left">8FA2</td></tr><tr><td>YEN SIGN</td><td align="left">00A5</td><td align="left">3F</td><td align="left">3F</td><td align="left">20</td><td align="left">3F</td></tr><tr><td>FULLWIDTH YEN SIGN</td><td align="left">FFE5</td><td align="left">818F</td><td align="left">818F</td><td align="left">A1EF</td><td align="left">3F</td></tr><tr><td>TILDE</td><td align="left">007E</td><td align="left">7E</td><td align="left">7E</td><td align="left">7E</td><td align="left">7E</td></tr><tr><td>OVERLINE</td><td align="left">203E</td><td align="left">3F</td><td align="left">3F</td><td align="left">20</td><td align="left">3F</td></tr><tr><td>HORIZONTAL BAR</td><td align="left">2015</td><td align="left">815C</td><td align="left">815C</td><td align="left">A1BD</td><td align="left">A1BD</td></tr><tr><td>EM DASH</td><td align="left">2014</td><td align="left">3F</td><td align="left">3F</td><td align="left">3F</td><td align="left">3F</td></tr><tr><td>REVERSE SOLIDUS</td><td align="left">005C</td><td align="left">815F</td><td align="left">5C</td><td align="left">5C</td><td align="left">5C</td></tr><tr><td>FULLWIDTH ""</td><td align="left">FF3C</td><td align="left">3F</td><td align="left">815F</td><td align="left">3F</td><td align="left">A1C0</td></tr><tr><td>WAVE DASH</td><td align="left">301C</td><td align="left">8160</td><td align="left">3F</td><td align="left">A1C1</td><td align="left">3F</td></tr><tr><td>FULLWIDTH TILDE</td><td align="left">FF5E</td><td align="left">3F</td><td align="left">8160</td><td align="left">3F</td><td align="left">A1C1</td></tr><tr><td>DOUBLE VERTICAL LINE</td><td align="left">2016</td><td align="left">8161</td><td align="left">3F</td><td align="left">A1C2</td><td align="left">3F</td></tr><tr><td>PARALLEL TO</td><td align="left">2225</td><td align="left">3F</td><td align="left">8161</td><td align="left">3F</td><td align="left">A1C2</td></tr><tr><td>MINUS SIGN</td><td align="left">2212</td><td align="left">817C</td><td align="left">3F</td><td align="left">A1DD</td><td align="left">3F</td></tr><tr><td>FULLWIDTH HYPHEN-MINUS</td><td align="left">FF0D</td><td align="left">3F</td><td align="left">817C</td><td align="left">3F</td><td align="left">A1DD</td></tr><tr><td>CENT SIGN</td><td align="left">00A2</td><td align="left">8191</td><td align="left">3F</td><td align="left">A1F1</td><td align="left">3F</td></tr><tr><td>FULLWIDTH CENT SIGN</td><td align="left">FFE0</td><td align="left">3F</td><td align="left">8191</td><td align="left">3F</td><td align="left">A1F1</td></tr><tr><td>POUND SIGN</td><td align="left">00A3</td><td align="left">8192</td><td align="left">3F</td><td align="left">A1F2</td><td align="left">3F</td></tr><tr><td>FULLWIDTH POUND SIGN</td><td align="left">FFE1</td><td align="left">3F</td><td align="left">8192</td><td align="left">3F</td><td align="left">A1F2</td></tr><tr><td>NOT SIGN</td><td align="left">00AC</td><td align="left">81CA</td><td align="left">3F</td><td align="left">A2CC</td><td align="left">3F</td></tr><tr><td>FULLWIDTH NOT SIGN</td><td align="left">FFE2</td><td align="left">3F</td><td align="left">81CA</td><td align="left">3F</td><td align="left">A2CC</td></tr></tbody></table></div><p> Now consider this portion of the table: </p><div class="informaltable"><table border="1"><colgroup><col><col align="right"><col align="right"><col align="right"></colgroup><thead><tr><th> </th><th align="right">ucs2</th><th align="right">sjis</th><th align="right">cp932</th></tr></thead><tbody><tr><td>NOT SIGN</td><td align="right">00AC</td><td align="right">81CA</td><td align="right">3F</td></tr><tr><td>FULLWIDTH NOT SIGN</td><td align="right">FFE2</td><td align="right">3F</td><td align="right">81CA</td></tr></tbody></table></div><p> This means that MySQL converts the <code class="literal">NOT SIGN</code> (Unicode <code class="literal">U+00AC</code>) to <code class="literal">sjis</code> code point <code class="literal">0x81CA</code> and to <code class="literal">cp932</code> code point <code class="literal">3F</code>. (<code class="literal">3F</code> is the question mark (“<span class="quote">?</span>”) — this is what is always used when the conversion cannot be performed. </p><p><a name="qandaitem-28-11-5"></a><span class="bold"><strong>28.11.5: </strong></span><span class="bold"><strong> What should I do if I want to convert SJIS <code class="literal">81CA</code> to <code class="literal">cp932</code>? </strong></span></p><p> Our answer is: “<span class="quote">?</span>”. There are serious complaints about this: many people would prefer a “<span class="quote">loose</span>” conversion, so that <code class="literal">81CA (NOT SIGN)</code> in <code class="literal">sjis</code> becomes <code class="literal">81CA (FULLWIDTH NOT SIGN)</code> in <code class="literal">cp932</code>. We are considering a change to this behavior. </p><p><a name="qandaitem-28-11-6"></a><span class="bold"><strong>28.11.6: </strong></span><span class="bold"><strong><a class="indexterm" name="id3171633"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3171648"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3171663"></a></strong></span><span class="bold"><strong> How does MySQL represent the Yen (<code class="literal">¥</code>) sign? </strong></span></p><p> A problem arises because some versions of Japanese character sets (both <code class="literal">sjis</code> and <code class="literal">euc</code>) treat <code class="literal">5C</code> as a <em class="firstterm">reverse solidus</em> (<code class="literal">\</code> — also known as a backslash), and others treat it as a yen sign (<code class="literal">¥</code>). </p><p> MySQL follows only one version of the JIS (Japanese Industrial Standards) standard description. In MySQL, <span class="emphasis"><em><code class="literal">5C</code> is always the reverse solidus (<code class="literal">\</code>)</em></span>. </p><p><a name="qandaitem-28-11-7"></a><span class="bold"><strong>28.11.7: </strong></span><span class="bold"><strong> Do MySQL plan to make a separate character set where <code class="literal">5C</code> is the Yen sign, as at least one other major DBMS does? </strong></span></p><p> This is one possible solution to the Yen sign issue; however, this will not happen in MySQL 5.1 or 5.2. </p><p><a name="qandaitem-28-11-8"></a><span class="bold"><strong>28.11.8: </strong></span><span class="bold"><strong><a class="indexterm" name="id3171784"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3171800"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3171815"></a></strong></span><span class="bold"><strong> Of what issues should I be aware when working with Korean character sets in MySQL? </strong></span></p><p> In theory, while there have been several versions of the <code class="literal">euckr</code> (<em class="firstterm">Extended Unix Code Korea</em>) character set, only one problem has been noted. </p><p> We use the “<span class="quote">ASCII</span>” variant of EUC-KR, in which the code point <code class="literal">0x5c</code> is REVERSE SOLIDUS, that is <code class="literal">\</code>, instead of the “<span class="quote">KS-Roman</span>” variant of EUC-KR, in which the code point <code class="literal">0x5c</code> is <code class="literal">WON SIGN</code>(<code class="literal">₩</code>). This means that you cannot convert Unicode <code class="literal">U+20A9</code> to <code class="literal">euckr</code>: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SELECT</code></strong> -> <strong class="userinput"><code>CONVERT('₩' USING euckr) AS euckr,</code></strong> -> <strong class="userinput"><code>HEX(CONVERT('₩' USING euckr)) AS hexeuckr;</code></strong> +-------+----------+ | euckr | hexeuckr | +-------+----------+ | ? | 3F | +-------+----------+ 1 row in set (0.00 sec) </pre><p> MySQL's graphic Korean chart is here: <a href="http://d.udm.net/bar/~bar/charts/euckr_korean_ci.html" target="_top">http://d.udm.net/bar/~bar/charts/euckr_korean_ci.html</a>. </p><p><a name="qandaitem-28-11-9"></a><span class="bold"><strong>28.11.9: </strong></span><span class="bold"><strong><a class="indexterm" name="id3171957"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3171973"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3171988"></a></strong></span><span class="bold"><strong> Why do I get <span class="errortext">Data truncated</span> error messages? </strong></span></p><p> For illustration, we'll create a table with one Unicode (<code class="literal">ucs2</code>) column and one Chinese (<code class="literal">gb2312</code>) column. </p><pre class="programlisting">mysql> <strong class="userinput"><code>CREATE TABLE ch</code></strong> -> <strong class="userinput"><code>(ucs2 CHAR(3) CHARACTER SET ucs2,</code></strong> -> <strong class="userinput"><code>gb2312 CHAR(3) CHARACTER SET gb2312);</code></strong> Query OK, 0 rows affected (0.05 sec) </pre><p> We'll try to place the rare character <code class="literal">汌</code> in both columns. </p><pre class="programlisting">mysql> <strong class="userinput"><code>INSERT INTO ch VALUES ('A汌B','A汌B');</code></strong> Query OK, 1 row affected, 1 warning (0.00 sec) </pre><p> Ah, there's a warning. Let's see what it is. </p><pre class="programlisting">mysql> <strong class="userinput"><code>SHOW WARNINGS;</code></strong> +---------+------+---------------------------------------------+ | Level | Code | Message | +---------+------+---------------------------------------------+ | Warning | 1265 | Data truncated for column 'gb2312' at row 1 | +---------+------+---------------------------------------------+ 1 row in set (0.00 sec) </pre><p> So it's a warning about the <code class="literal">gb2312</code> column only. </p><pre class="programlisting">mysql> SELECT ucs2,HEX(ucs2),gb2312,HEX(gb2312) FROM ch; +-------+--------------+--------+-------------+ | ucs2 | HEX(ucs2) | gb2312 | HEX(gb2312) | +-------+--------------+--------+-------------+ | A汌B | 00416C4C0042 | A?B | 413F42 | +-------+--------------+--------+-------------+ 1 row in set (0.00 sec) </pre><p> There are several things that need explanation here. </p><div class="orderedlist"><ol type="1"><li><p> The fact that it's a “<span class="quote">warning</span>” rather than an “<span class="quote">error</span>” is characteristic of MySQL. We like to try to do what we can, to get the best fit, rather than give up. </p></li><li><p> The <code class="literal">汌</code> character isn't in the <code class="literal">gb2312</code> character set. We described that problem earlier. </p></li><li><p> Admittedly the message is misleading. We didn't “<span class="quote">truncate</span>” in this case, we replaced with a question mark. We've had a complaint about this message (See <a href="http://bugs.mysql.com/9337" target="_top">Bug#9337</a>). But until we come up with something better, just accept that error/warning code 2165 can mean a variety of things. </p></li><li><p> With <code class="literal">SQL_MODE=TRADITIONAL</code>, there would be an error message, but instead of error 2165 you would see: <code class="literal">ERROR 1406 (22001): Data too long for column 'gb2312' at row 1</code>. </p></li></ol></div><p> </p><p><a name="qandaitem-28-11-10"></a><span class="bold"><strong>28.11.10: </strong></span><span class="bold"><strong><a class="indexterm" name="id3172214"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3172229"></a></strong></span><span class="bold"><strong> Why does my GUI front end or browser not display CJK characters correctly in my application using Access, PHP, or another API? </strong></span></p><p> Obtain a direct connection to the server using the <span><strong class="command">mysql</strong></span> client (Windows: <span><strong class="command">mysql.exe</strong></span>), and try the same query there. If <span><strong class="command">mysql</strong></span> responds correctly, then the trouble may be that your application interface requires initialization. Use <span><strong class="command">mysql</strong></span> to tell you what character set or sets it uses with the statement <code class="literal">SHOW VARIABLES LIKE 'char%';</code>. If you are using Access, then you are most likely connecting with MyODBC. In this case, you should check <a href="connectors.html#myodbc-configuration" title="24.1.3. Connector/ODBC Configuration">Section 24.1.3, “Connector/ODBC Configuration”</a>. If, for instance, you use <code class="literal">big5</code>, you would enter <code class="literal">SET NAMES 'big5'</code>. (Note that no <code class="literal">;</code> is required in this case). If you are using ASP, you might need to add <code class="literal">SET NAMES</code> in the code. Here is an example that has worked in the past: </p><pre class="programlisting"><% Session.CodePage=0 Dim strConnection Dim Conn strConnection="driver={MySQL ODBC 3.51 Driver};server=<em class="replaceable"><code>server</code></em>;uid=<em class="replaceable"><code>username</code></em>;" \ & "pwd=<em class="replaceable"><code>password</code></em>;database=<em class="replaceable"><code>database</code></em>;stmt=SET NAMES 'big5';" Set Conn = Server.CreateObject("ADODB.Connection") Conn.Open strConnection %> </pre><p> In much the same way, if you are using any character set other than <code class="literal">latin1</code> with Connector/NET, then you must specify the character set in the connection string. See <a href="connectors.html#connector-net-using-connecting" title="24.2.5.1. Connecting to MySQL Using Connector/NET">Section 24.2.5.1, “Connecting to MySQL Using Connector/NET”</a>, for more information. </p><p> If you are using PHP, try this: </p><pre class="programlisting"><?php $link = mysql_connect($host, $usr, $pwd); mysql_select_db($db); if( mysql_error() ) { print "Database ERROR: " . mysql_error(); } mysql_query("SET NAMES 'utf8'", $link); ?> </pre><p> In this case, we used <code class="literal">SET NAMES</code> to change <code class="literal">character_set_client</code> and <code class="literal">character_set_connection</code> and <code class="literal">character_set_results</code>. </p><p> We encourage the use of the newer <code class="literal">mysqli</code> extension, rather than <code class="literal">mysql</code>. Using <code class="literal">mysqli</code>, the previous example could be rewritten as shown here: </p><pre class="programlisting"> <?php $link = new mysqli($host, $usr, $pwd, $db); if( mysqli_connect_errno() ) { printf("Connect failed: %s\n", mysqli_connect_error()); exit(); } $link->query("SET NAMES 'utf8'"); ?> </pre><p> Another issue often encountered in PHP applications has to do with assumptions made by the browser. Sometimes adding or changing a <code class="literal"><meta></code> tag suffices to correct the problem: for example, to insure that the user agent interprets page content as <code class="literal">UTF-8</code>, you should include <code class="literal"><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></code> in the <code class="literal"><head></code> of the HTML page. </p><p> If you are using Connector/J, see <a href="connectors.html#connector-j-reference-charsets" title="24.4.4.4. Using Character Sets and Unicode">Section 24.4.4.4, “Using Character Sets and Unicode”</a>. </p><p><a name="qandaitem-28-11-11"></a><span class="bold"><strong>28.11.11: </strong></span><span class="bold"><strong><a class="indexterm" name="id3172501"></a></strong></span><span class="bold"><strong> I've upgraded to MySQL 5.0. How can I revert to behavior like that in MySQL 4.0 with regard to character sets? </strong></span></p><p> In MySQL Version 4.0, there was a single “<span class="quote">global</span>” character set for both server and client, and the decision as to which character to use was made by the server administrator. This changed starting with MySQL Version 4.1. What happens now is a “<span class="quote">handshake</span>”, as described in <a href="internationalization-localization.html#charset-connection" title="9.1.4. Connection Character Sets and Collations">Section 9.1.4, “Connection Character Sets and Collations”</a>: </p><div class="blockquote"><blockquote class="blockquote"><p> When a client connects, it sends to the server the name of the character set that it wants to use. The server uses the name to set the <code class="literal">character_set_client</code>, <code class="literal">character_set_results</code>, and <code class="literal">character_set_connection</code> system variables. In effect, the server performs a <code class="literal">SET NAMES</code> operation using the character set name. </p></blockquote></div><p> The effect of this is that you cannot control the client character set by starting <span><strong class="command">mysqld</strong></span> with <code class="option">--character-set-server=utf8</code>. However, some of our Asian customers have said that prefer the MySQL 4.0 behavior. To make it possible to retain this behavior, we added a <span><strong class="command">mysqld</strong></span> switch, <code class="option">--character-set-client-handshake</code>, which can be turned off with <code class="option">--skip-character-set-client-handshake</code>. If you start <span><strong class="command">mysqld</strong></span> with <code class="option">--skip-character-set-client-handshake</code>, then, when a client connects, it sends to the server the name of the character set that it wants to use — however, <span class="emphasis"><em>the server ignores this request from the client</em></span>. </p><p> By way of example, suppose that your favorite server character set is <code class="literal">latin1</code> (unlikely in a CJK area, but this is the default value). Suppose further that the client uses <code class="literal">utf8</code> because this is what the client's operating system supports. Now, start the server with <code class="literal">latin1</code> as its default character set: </p><pre class="programlisting">mysqld --character-set-server=latin1 </pre><p> And then start the client with the default character set <code class="literal">utf8</code>: </p><pre class="programlisting">mysql --default-character-set=utf8 </pre><p> The current settings can be seen by viewing the output of <code class="literal">SHOW VARIABLES</code>: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SHOW VARIABLES LIKE 'char%';</code></strong> +--------------------------+----------------------------------------+ | Variable_name | Value | +--------------------------+----------------------------------------+ | character_set_client | utf8 | | character_set_connection | utf8 | | character_set_database | latin1 | | character_set_filesystem | binary | | character_set_results | utf8 | | character_set_server | latin1 | | character_set_system | utf8 | | character_sets_dir | /usr/local/mysql/share/mysql/charsets/ | +--------------------------+----------------------------------------+ 8 rows in set (0.01 sec) </pre><p> Now stop the client, and then stop the server using <span><strong class="command">mysqladmin</strong></span>. Then start the server again, but this time tell it to skip the handshake like so: </p><pre class="programlisting">mysqld --character-set-server=utf8 --skip-character-set-client-handshake </pre><p> Start the client with <code class="literal">utf8</code> once again as the default character set, then display the current settings: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SHOW VARIABLES LIKE 'char%';</code></strong> +--------------------------+----------------------------------------+ | Variable_name | Value | +--------------------------+----------------------------------------+ | character_set_client | latin1 | | character_set_connection | latin1 | | character_set_database | latin1 | | character_set_filesystem | binary | | character_set_results | latin1 | | character_set_server | latin1 | | character_set_system | utf8 | | character_sets_dir | /usr/local/mysql/share/mysql/charsets/ | +--------------------------+----------------------------------------+ 8 rows in set (0.01 sec) </pre><p> As you can see by comparing the differing results from <code class="literal">SHOW VARIABLES</code>, the server ignores the client's initial settings if the <code class="option">--skip-character-set-client-handshake</code> is used. </p><p><a name="qandaitem-28-11-12"></a><span class="bold"><strong>28.11.12: </strong></span><span class="bold"><strong><a class="indexterm" name="id3172802"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3172817"></a></strong></span><span class="bold"><strong> Why do some <code class="literal">LIKE</code> and <code class="literal">FULLTEXT</code> searches with CJK characters fail? </strong></span></p><p> There is a very simple problem with <code class="literal">LIKE</code> searches on <code class="literal">BINARY</code> and <code class="literal">BLOB</code> columns: we need to know the end of a character. With multi-byte character sets, different characters might have different octet lengths. For example, in <code class="literal">utf8</code>, <code class="literal">A</code> requires one byte but <code class="literal">ペ</code> requires three bytes, as shown here: </p><pre class="programlisting">+-------------------------+---------------------------+ | OCTET_LENGTH(_utf8 'A') | OCTET_LENGTH(_utf8 'ペ') | +-------------------------+---------------------------+ | 1 | 3 | +-------------------------+---------------------------+ 1 row in set (0.00 sec) </pre><p> If we don't know where the first character ends, then we don't know where the second character begins, in which case even very simple searches such as <code class="literal">LIKE '_A%'</code> fail. The solution is to use a regular CJK character set in the first place, or to convert to a CJK character set before comparing. </p><p> This is one reason why MySQL cannot allow encodings of nonexistent characters. If it is not strict about rejecting bad input, then it has no way of knowing where characters end. </p><p> For <code class="literal">FULLTEXT</code> searches, we need to know where words begin and end. With Western languages, this is rarely a problem because most (if not all) of these use an easy-to-identify word boundary — the space character. However, this is not usually the case with Asian writing. We could use arbitrary halfway measures, like assuming that all Han characters represent words, or (for Japanese) depending on changes from Katakana to Hiragana due to grammatical endings. However, the only sure solution requires a comprehensive word list, which means that we would have to include a dictionary in the server for each Asian language supported. This is simply not feasible. </p><p><a name="qandaitem-28-11-13"></a><span class="bold"><strong>28.11.13: </strong></span><span class="bold"><strong><a class="indexterm" name="id3172966"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3172981"></a></strong></span><span class="bold"><strong> What CJK character sets are available in MySQL? </strong></span></p><p> The list of CJK character sets may vary depending on your MySQL version. For example, the <code class="literal">eucjpms</code> character set was not supported prior to MySQL 5.0.3 (see <a href="news.html#news-5-0-3" title="E.1.27. Changes in MySQL 5.0.3 (23 March 2005: Beta)">Section E.1.27, “Changes in MySQL 5.0.3 (23 March 2005: Beta)”</a>). However, since the name of the applicable language appears in the <code class="literal">DESCRIPTION</code> column for every entry in the <code class="literal">INFORMATION_SCHEMA.CHARACTER_SETS</code> table, you can obtain a current list of all the non-Unicode CJK character sets using this query: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SELECT CHARACTER_SET_NAME, DESCRIPTION</code></strong> -> <strong class="userinput"><code>FROM INFORMATION_SCHEMA.CHARACTER_SETS</code></strong> -> <strong class="userinput"><code>WHERE DESCRIPTION LIKE '%Chinese%'</code></strong> -> <strong class="userinput"><code>OR DESCRIPTION LIKE '%Japanese%'</code></strong> -> <strong class="userinput"><code>OR DESCRIPTION LIKE '%Korean%'</code></strong> -> <strong class="userinput"><code>ORDER BY CHARACTER_SET_NAME;</code></strong> +--------------------+---------------------------+ | CHARACTER_SET_NAME | DESCRIPTION | +--------------------+---------------------------+ | big5 | Big5 Traditional Chinese | | cp932 | SJIS for Windows Japanese | | eucjpms | UJIS for Windows Japanese | | euckr | EUC-KR Korean | | gb2312 | GB2312 Simplified Chinese | | gbk | GBK Simplified Chinese | | sjis | Shift-JIS Japanese | | ujis | EUC-JP Japanese | +--------------------+---------------------------+ 8 rows in set (0.01 sec) </pre><p> (See <a href="information-schema.html#character-sets-table" title="21.9. The INFORMATION_SCHEMA CHARACTER_SETS Table">Section 21.9, “The <code class="literal">INFORMATION_SCHEMA CHARACTER_SETS</code> Table”</a>, for more information.) </p><p><a name="qandaitem-28-11-14"></a><span class="bold"><strong>28.11.14: </strong></span><span class="bold"><strong><a class="indexterm" name="id3173107"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173122"></a></strong></span><span class="bold"><strong> How do I know whether character <em class="replaceable"><code>X</code></em> is available in all character sets? </strong></span></p><p> The majority of simplified Chinese and basic non-halfwidth Japanese <span class="foreignphrase"><em class="foreignphrase">Kana</em></span> characters appear in all CJK character sets. This stored procedure accepts a <code class="literal">UCS-2</code> Unicode character, converts it to all other character sets, and displays the results in hexadecimal. </p><pre class="programlisting">DELIMITER // CREATE PROCEDURE p_convert(ucs2_char CHAR(1) CHARACTER SET ucs2) BEGIN CREATE TABLE tj (ucs2 CHAR(1) character set ucs2, utf8 CHAR(1) character set utf8, big5 CHAR(1) character set big5, cp932 CHAR(1) character set cp932, eucjpms CHAR(1) character set eucjpms, euckr CHAR(1) character set euckr, gb2312 CHAR(1) character set gb2312, gbk CHAR(1) character set gbk, sjis CHAR(1) character set sjis, ujis CHAR(1) character set ujis); INSERT INTO tj (ucs2) VALUES (ucs2_char); UPDATE tj SET utf8=ucs2, big5=ucs2, cp932=ucs2, eucjpms=ucs2, euckr=ucs2, gb2312=ucs2, gbk=ucs2, sjis=ucs2, ujis=ucs2; /* If there's a conversion problem, UPDATE will produce a warning. */ SELECT hex(ucs2) AS ucs2, hex(utf8) AS utf8, hex(big5) AS big5, hex(cp932) AS cp932, hex(eucjpms) AS eucjpms, hex(euckr) AS euckr, hex(gb2312) AS gb2312, hex(gbk) AS gbk, hex(sjis) AS sjis, hex(ujis) AS ujis FROM tj; DROP TABLE tj; END// </pre><p> The input can be any single <code class="literal">ucs2</code> character, or it can be the code point value (hexadecimal representation) of that character. For example, from Unicode's list of <code class="literal">ucs2</code> encodings and names (<a href="http://www.unicode.org/Public/UNIDATA/UnicodeData.txt" target="_top">http://www.unicode.org/Public/UNIDATA/UnicodeData.txt</a>), we know that the <span class="foreignphrase"><em class="foreignphrase">Katakana</em></span> character <span class="foreignphrase"><em class="foreignphrase">Pe</em></span> appears in all CJK character sets, and that its code point value is <code class="literal">0x30da</code>. If we use this value as the argument to <code class="literal">p_convert()</code>, the result is as shown here: </p><pre class="programlisting">mysql> <strong class="userinput"><code>CALL p_convert(0x30da)//</code></strong> +------+--------+------+-------+---------+-------+--------+------+------+------+ | ucs2 | utf8 | big5 | cp932 | eucjpms | euckr | gb2312 | gbk | sjis | ujis | +------+--------+------+-------+---------+-------+--------+------+------+------+ | 30DA | E3839A | C772 | 8379 | A5DA | ABDA | A5DA | A5DA | 8379 | A5DA | +------+--------+------+-------+---------+-------+--------+------+------+------+ 1 row in set (0.04 sec) </pre><p> Since none of the column values is <code class="literal">3F</code> — that is, the question mark character (<code class="literal">?</code>) — we know that every conversion worked. </p><p><a name="qandaitem-28-11-15"></a><span class="bold"><strong>28.11.15: </strong></span><span class="bold"><strong><a class="indexterm" name="id3173281"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173296"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173312"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173327"></a></strong></span><span class="bold"><strong> Why don't CJK strings sort correctly in Unicode? (I) </strong></span></p><p> Sometimes people observe that the result of a <code class="literal">utf8_unicode_ci</code> or <code class="literal">ucs2_unicode_ci</code> search, or of an <code class="literal">ORDER BY</code> sort is not what they think a native would expect. Although we never rule out the possibility that there is a bug, we have found in the past that many people do not read correctly the standard table of weights for the Unicode Collation Algorithm. MySQL uses the table found at <a href="http://www.unicode.org/Public/UCA/4.0.0/allkeys-4.0.0.txt" target="_top">http://www.unicode.org/Public/UCA/4.0.0/allkeys-4.0.0.txt</a>. This is not the first table you will find by navigating from the <code class="literal">unicode.org</code> home page, because MySQL uses the older 4.0.0 “<span class="quote">allkeys</span>” table, rather than the more recent 4.1.0 table. This is because we are very wary about changing ordering which affects indexes, lest we bring about situations such as that reported in <a href="http://bugs.mysql.com/16526" target="_top">Bug#16526</a>, illustrated as follows: </p><pre class="programlisting">mysql< <strong class="userinput"><code>CREATE TABLE tj (s1 CHAR(1) CHARACTER SET utf8 COLLATE utf8_unicode_ci);</code></strong> Query OK, 0 rows affected (0.05 sec) mysql> <strong class="userinput"><code>INSERT INTO tj VALUES ('が'),('か');</code></strong> Query OK, 2 rows affected (0.00 sec) Records: 2 Duplicates: 0 Warnings: 0 mysql> <strong class="userinput"><code>SELECT * FROM tj WHERE s1 = 'か';</code></strong> +------+ | s1 | +------+ | が | | か | +------+ 2 rows in set (0.00 sec) </pre><p> The character in the first result row is not the one that we searched for. Why did MySQL retrieve it? First we look for the Unicode code point value, which is possible by reading the hexadecimal number for the <code class="literal">ucs2</code> version of the characters: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SELECT s1, HEX(CONVERT(s1 USING ucs2)) FROM tj;</code></strong> +------+-----------------------------+ | s1 | HEX(CONVERT(s1 USING ucs2)) | +------+-----------------------------+ | が | 304C | | か | 304B | +------+-----------------------------+ 2 rows in set (0.03 sec) </pre><p> Now we search for <code class="literal">304B</code> and <code class="literal">304C</code> in the <code class="literal">4.0.0 allkeys</code> table, and find these lines: </p><pre class="programlisting">304B ; [.1E57.0020.000E.304B] # HIRAGANA LETTER KA 304C ; [.1E57.0020.000E.304B][.0000.0140.0002.3099] # HIRAGANA LETTER GA; QQCM </pre><p> The official Unicode names (following the “<span class="quote">#</span>” mark) tell us the Japanese syllabary (Hiragana), the informal classification (letter, digit, or punctuation mark), and the Western identifier (<code class="literal">KA</code> or <code class="literal">GA</code>, which happen to be voiced and unvoiced components of the same letter pair). More importantly, the <em class="firstterm">primary weight</em> (the first hexadecimal number inside the square brackets) is <code class="literal">1E57</code> on both lines. For comparisons in both searching and sorting, MySQL pays attention to the primary weight only, ignoring all the other numbers. This means that we are sorting <code class="literal">が</code> and <code class="literal">か</code> correctly according to the Unicode specification. If we wanted to distinguish them, we'd have to use a non-UCA (Unicode Collation Algorithm) collation (<code class="literal">utf8_unicode_bin</code> or <code class="literal">utf8_general_ci</code>), or to compare the <code class="literal">HEX()</code> values, or use <code class="literal">ORDER BY CONVERT(s1 USING sjis)</code>. Being correct “<span class="quote">according to Unicode</span>” isn't enough, of course: the person who submitted the bug was equally correct. We plan to add another collation for Japanese according to the JIS X 4061 standard, in which voiced/unvoiced letter pairs like <code class="literal">KA</code>/<code class="literal">GA</code> are distinguishable for ordering purposes. </p><p><a name="qandaitem-28-11-16"></a><span class="bold"><strong>28.11.16: </strong></span><span class="bold"><strong><a class="indexterm" name="id3173602"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173617"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173633"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173648"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173663"></a></strong></span><span class="bold"><strong> Why don't CJK strings sort correctly in Unicode? (II) </strong></span></p><p> If you are using Unicode (<code class="literal">ucs2</code> or <code class="literal">utf8</code>), and you know what the Unicode sort order is (see <a href="faqs.html#faqs-cjk" title="A.11. MySQL 5.0 FAQ — MySQL Chinese, Japanese, and Korean Character Sets">Section A.11, “MySQL 5.0 FAQ — MySQL Chinese, Japanese, and Korean Character Sets”</a>), but MySQL still seems to sort your table incorrectly, then you should first verify the table character set: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SHOW CREATE TABLE t\G</code></strong> ******************** 1. row ****************** Table: t Create Table: CREATE TABLE `t` ( `s1` char(1) CHARACTER SET ucs2 DEFAULT NULL ) ENGINE=MyISAM DEFAULT CHARSET=latin1 1 row in set (0.00 sec) </pre><p> Since the character set appears to be correct, let's see what information the <code class="literal">INFORMATION_SCHEMA.COLUMNS</code> table can provide about this column: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SELECT COLUMN_NAME, CHARACTER_SET_NAME, COLLATION_NAME</code></strong> -> <strong class="userinput"><code>FROM INFORMATION_SCHEMA.COLUMNS</code></strong> -> <strong class="userinput"><code>WHERE COLUMN_NAME = 's1'</code></strong> -> <strong class="userinput"><code>AND TABLE_NAME = 't';</code></strong> +-------------+--------------------+-----------------+ | COLUMN_NAME | CHARACTER_SET_NAME | COLLATION_NAME | +-------------+--------------------+-----------------+ | s1 | ucs2 | ucs2_general_ci | +-------------+--------------------+-----------------+ 1 row in set (0.01 sec) </pre><p> (See <a href="information-schema.html#columns-table" title="21.3. The INFORMATION_SCHEMA COLUMNS Table">Section 21.3, “The <code class="literal">INFORMATION_SCHEMA COLUMNS</code> Table”</a>, for more information.) </p><p> You can see that the collation is <code class="literal">ucs2_general_ci</code> instead of <code class="literal">ucs2_unicode_ci</code>. The reason why this is so can be found using <code class="literal">SHOW CHARSET</code>, as shown here: </p><pre class="programlisting">mysql> <strong class="userinput"><code>SHOW CHARSET LIKE 'ucs2%';</code></strong> +---------+---------------+-------------------+--------+ | Charset | Description | Default collation | Maxlen | +---------+---------------+-------------------+--------+ | ucs2 | UCS-2 Unicode | ucs2_general_ci | 2 | +---------+---------------+-------------------+--------+ 1 row in set (0.00 sec) </pre><p> For <code class="literal">ucs2</code> and <code class="literal">utf8</code>, the default collation is “<span class="quote">general</span>”. To specify a Unicode collation, use <code class="literal">COLLATE ucs2_unicode_ci</code>. </p><p><a name="qandaitem-28-11-17"></a><span class="bold"><strong>28.11.17: </strong></span><span class="bold"><strong><a class="indexterm" name="id3173850"></a></strong></span><span class="bold"><strong> Why are my supplementary characters rejected by MySQL? </strong></span></p><p> MySQL does not support supplementary characters — that is, characters which need more than 3 bytes — for <code class="literal">UTF-8</code>. We support only what Unicode calls the <span class="emphasis"><em>Basic Multilingual Plane / Plane 0</em></span>. Only a few very rare Han characters are supplementary; support for them is uncommon. This has led to reports such as that found in <a href="http://bugs.mysql.com/12600" target="_top">Bug#12600</a>, which we rejected as “<span class="quote">not a bug</span>”. With <code class="literal">utf8</code>, we must truncate an input string when we encounter bytes that we don't understand. Otherwise, we wouldn't know how long the bad multi-byte character is. </p><p> One possible workaround is to use <code class="literal">ucs2</code> instead of <code class="literal">utf8</code>, in which case the “<span class="quote">bad</span>” characters are changed to question marks; however, no truncation takes place. You can also change the data type to <code class="literal">BLOB</code> or <code class="literal">BINARY</code>, which perform no validity checking. </p><p> We intend at some point in the future to add support for <code class="literal">UTF-16</code>, which would solve such issues by allowing 4-byte characters. However, we have as yet set no definite timetable for doing so. </p><p><a name="qandaitem-28-11-18"></a><span class="bold"><strong>28.11.18: </strong></span><span class="bold"><strong><a class="indexterm" name="id3173970"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3173985"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3174000"></a></strong></span><span class="bold"><strong> Shouldn't it be “<span class="quote">CJKV</span>”? </strong></span></p><p> No. The term “<span class="quote">CJKV</span>” (<em class="firstterm">Chinese Japanese Korean Vietnamese</em>) refers to Vietnamese character sets which contain Han (originally Chinese) characters. MySQL has no plan to support the old Vietnamese script using Han characters. MySQL does of course support the modern Vietnamese script with Western characters. </p><p> <a href="http://bugs.mysql.com/4745" target="_top">Bug#4745</a> is a request for a specialized Vietnamese collation, which we might add in the future if there is sufficient demand for it. </p><p><a name="qandaitem-28-11-19"></a><span class="bold"><strong>28.11.19: </strong></span><span class="bold"><strong><a class="indexterm" name="id3174058"></a></strong></span><span class="bold"><strong> Does MySQL allow CJK characters to be used in database and table names? </strong></span></p><p> This issue is fixed in MySQL 5.1, by automatically rewriting the names of the corresponding directories and files. </p><p> For example, if you create a database named <code class="literal">楮</code> on a server whose operating system does not support CJK in directory names, MySQL creates a directory named <code class="literal">@0w@00a5@00ae</code>. which is just a fancy way of encoding <code class="literal">E6A5AE</code> — that is, the Unicode hexadecimal representation for the <code class="literal">楮</code> character. However, if you run a <code class="literal">SHOW DATABASES</code> statement, you can see that the database is listed as <code class="literal">楮</code>. </p><p><a name="qandaitem-28-11-20"></a><span class="bold"><strong>28.11.20: </strong></span><span class="bold"><strong><a class="indexterm" name="id3174143"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3174158"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3174174"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3174189"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3174204"></a></strong></span><span class="bold"><strong><a class="indexterm" name="id3174219"></a></strong></span><span class="bold"><strong></strong></span><span class="bold"><strong> Where can I find translations of the MySQL Manual into Chinese, Japanese, and Korean? </strong></span></p><p> A Simplified Chinese version of the Manual, current for MySQL 5.1.12, can be found at <a href="http://dev.mysql.com/doc/#chinese-5.1" target="_top">http://dev.mysql.com/doc/#chinese-5.1</a>. The Japanese translation of the MySQL 4.1 manual can be downloaded from <a href="http://dev.mysql.com/doc/#japanese-4.1" target="_top">http://dev.mysql.com/doc/#japanese-4.1</a>. </p><p><a name="qandaitem-28-11-21"></a><span class="bold"><strong>28.11.21: </strong></span><span class="bold"><strong> Where can I get help with CJK and related issues in MySQL? </strong></span></p><p> The following resources are available: </p><div class="itemizedlist"><ul type="disc"><li><p> A listing of MySQL user groups can be found at <a href="http://dev.mysql.com/user-groups/" target="_top">http://dev.mysql.com/user-groups/</a>. </p></li><li><p> You can contact a sales engineer at the MySQL KK Japan office using any of the following: </p><pre class="programlisting">Tel: +81(0)3-5326-3133 Fax: +81(0)3-5326-3001 Email: dsaito@mysql.com </pre><p> </p></li><li><p> View feature requests relating to character set issues at <a href="http://tinyurl.com/y6xcuf" target="_top">http://tinyurl.com/y6xcuf</a>. </p></li><li><p> Visit the MySQL <a href="http://forums.mysql.com/list.php?103" target="_top">Character Sets, Collation, Unicode Forum</a>. We are also in the process of adding foreign-language forums at <a href="http://forums.mysql.com/" target="_top">http://forums.mysql.com/</a>. </p></li></ul></div><p> </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-connectors-apis"></a>A.12. MySQL 5.0 FAQ — Connectors & APIs</h2></div></div></div><p> For common questions, issues, and answers relating to the MySQL Connectors and other APIs, see the following areas of the Manual: </p><div class="itemizedlist"><ul type="disc"><li><p> <a href="apis.html#c-api-problems" title="23.2.14. Common Questions and Problems When Using the C API">Section 23.2.14, “Common Questions and Problems When Using the C API”</a> </p></li><li><p> <a href="apis.html#php-problems" title="23.3.1. Common Problems with MySQL and PHP">Section 23.3.1, “Common Problems with MySQL and PHP”</a> </p></li><li><p> <a href="connectors.html#myodbc-usagenotes" title="24.1.6. Connector/ODBC Notes and Tips">Section 24.1.6, “Connector/ODBC Notes and Tips”</a> </p></li><li><p> <a href="connectors.html#connector-net-using" title="24.2.5. Connector/NET Notes and Tips">Section 24.2.5, “Connector/NET Notes and Tips”</a> </p></li><li><p> <a href="connectors.html#connector-j-usagenotes" title="24.4.5. Connector/J Notes and Tips">Section 24.4.5, “Connector/J Notes and Tips”</a> </p></li><li><p> <a href="connectors.html#connector-mxj-usagenotes" title="24.5.5. Connector/MXJ Notes and Tips">Section 24.5.5, “Connector/MXJ Notes and Tips”</a> </p></li></ul></div></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-replication"></a>A.13. MySQL 5.0 FAQ — Replication</h2></div></div></div><p> For answers to common queries and question regarding Replication within MySQL, see <a href="replication.html#replication-faq" title="15.3.4. Replication FAQ">Section 15.3.4, “Replication FAQ”</a>. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a name="faqs-mysql-drbd-heartbeat"></a>A.14. MySQL 5.0 FAQ — MySQL, DRBD, and Heartbeat</h2></div></div></div><div class="toc"><dl><dt><span class="section"><a href="faqs.html#faqs-drbd">A.14.1. Distributed Replicated Block Device</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-linux-heartbeat">A.14.2. Linux Heartbeat</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-architecture">A.14.3. DRBD Architecture</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-mysql-replication-scale">A.14.4. DRBD and MySQL Replication</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-file-systems">A.14.5. DRBD and File Systems</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-lvm">A.14.6. DRBD and LVM</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-virtualization">A.14.7. DRBD and Virtualization</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-security">A.14.8. DRBD and Security</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-system-requirements">A.14.9. DRBD and System Requirements</a></span></dt><dt><span class="section"><a href="faqs.html#drbd-support-consulting">A.14.10. DBRD and Support and Consulting</a></span></dt></dl></div><a class="indexterm" name="id3174466"></a><a class="indexterm" name="id3174478"></a><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="faqs-drbd"></a>A.14.1. Distributed Replicated Block Device</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about Distributed Replicated Block Device (DRBD). </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-1-1">28.14.1.1: </a><a class="indexterm" name="id3174515"></a> What is DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-1-2">28.14.1.2: </a> What are “<span class="quote">Block Devices</span>”? </p></li><li><p><a href="faqs.html#qandaitem-28-14-1-3">28.14.1.3: </a><a class="indexterm" name="id3174544"></a> How is DRBD licensed? </p></li><li><p><a href="faqs.html#qandaitem-28-14-1-4">28.14.1.4: </a> Where can I download DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-1-5">28.14.1.5: </a> If I find a bug in DRBD, to whom do I submit the issue? </p></li><li><p><a href="faqs.html#qandaitem-28-14-1-6">28.14.1.6: </a> Where can I get more technical and business information concerning MySQL and DRBD? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-1-1"></a><span class="bold"><strong>28.14.1.1: </strong></span><span class="bold"><strong><a class="indexterm" name="id3174601"></a></strong></span><span class="bold"><strong> What is DRBD? </strong></span></p><p> DRBD is an acronym for Distributed Replicated Block Device. DRBD is an open source Linux kernel block device which leverages synchronous replication to achieve a consistent view of data between two systems, typically an Active and Passive system. DRBD currently supports all the major flavors of Linux and comes bundled in several major Linux distributions. The DRBD project is maintained by <a href="http://www.drbd.org/" target="_top">LINBIT</a>. </p><p><a name="qandaitem-28-14-1-2"></a><span class="bold"><strong>28.14.1.2: </strong></span><span class="bold"><strong> What are “<span class="quote">Block Devices</span>”? </strong></span></p><p> Block devices are the type of device used to represent storage in the Linux Kernel. All physical disk devices present a “<span class="quote">block device</span>” interface. Additionally, virtual disk systems like LVM or DRBD present a “<span class="quote">block device</span>” interface. In this way, the file system or other software that might want to access a disk device can be used with any number of real or virtual devices without having to know anything about their underlying implementation details. </p><p><a name="qandaitem-28-14-1-3"></a><span class="bold"><strong>28.14.1.3: </strong></span><span class="bold"><strong><a class="indexterm" name="id3174683"></a></strong></span><span class="bold"><strong> How is DRBD licensed? </strong></span></p><p> DRBD is licensed under the GPL. </p><p><a name="qandaitem-28-14-1-4"></a><span class="bold"><strong>28.14.1.4: </strong></span><span class="bold"><strong> Where can I download DRBD? </strong></span></p><p> Please see <a href="http://www.drbd.org/download.html" target="_top">http://www.drbd.org/download.html</a> </p><p><a name="qandaitem-28-14-1-5"></a><span class="bold"><strong>28.14.1.5: </strong></span><span class="bold"><strong> If I find a bug in DRBD, to whom do I submit the issue? </strong></span></p><p> Bug reports should be submitted to the DRBD mailing list. Please see: <a href="http://lists.linbit.com/%20" target="_top">http://lists.linbit.com/ </a>. </p><p><a name="qandaitem-28-14-1-6"></a><span class="bold"><strong>28.14.1.6: </strong></span><span class="bold"><strong> Where can I get more technical and business information concerning MySQL and DRBD? </strong></span></p><p> Please visit: <a href="http://mysql.com/drbd/" target="_top">http://mysql.com/drbd/</a> </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-linux-heartbeat"></a>A.14.2. Linux Heartbeat</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about Linux Heartbeat. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-2-1">28.14.2.1: </a> What is Linux Heartbeat? </p></li><li><p><a href="faqs.html#qandaitem-28-14-2-2">28.14.2.2: </a> How is Linux Heartbeat licensed? </p></li><li><p><a href="faqs.html#qandaitem-28-14-2-3">28.14.2.3: </a> Where can I download Linux Heartbeat? </p></li><li><p><a href="faqs.html#qandaitem-28-14-2-4">28.14.2.4: </a> If I find a bug with Linux Heartbeat, to whom do I submit the issue? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-2-1"></a><span class="bold"><strong>28.14.2.1: </strong></span><span class="bold"><strong> What is Linux Heartbeat? </strong></span></p><p> The Linux-HA project (<a href="http://www.linux-ha.org/" target="_top">http://www.linux-ha.org/</a>) offers a high availability solution commonly referred to as Linux Heartbeat. Linux Heartbeat ships as part of several Linux distributions, as well as within several embedded high availability systems. This solution can also be used for other applications besides databases servers, such as mail servers, web servers, file servers, and DNS servers. </p><p> Linux Heartbeat implements a heartbeat-protocol. A heartbeat-protocol means that messages are sent at regular intervals between two or more nodes. If a message is not received from a node within a given interval, then it is assumed the node has failed and some type of failover or recovery action is required. Linux Heartbeat is typically configured to send these heartbeat messages over standard Ethernet interfaces, but it does also support other methods, such as serial-line links. </p><p><a name="qandaitem-28-14-2-2"></a><span class="bold"><strong>28.14.2.2: </strong></span><span class="bold"><strong> How is Linux Heartbeat licensed? </strong></span></p><p> Linux Heartbeat is licensed under the GPL. </p><p><a name="qandaitem-28-14-2-3"></a><span class="bold"><strong>28.14.2.3: </strong></span><span class="bold"><strong> Where can I download Linux Heartbeat? </strong></span></p><p> Please see <a href="http://linux-ha.org/download/index.html" target="_top">http://linux-ha.org/download/index.html</a>. </p><p><a name="qandaitem-28-14-2-4"></a><span class="bold"><strong>28.14.2.4: </strong></span><span class="bold"><strong> If I find a bug with Linux Heartbeat, to whom do I submit the issue? </strong></span></p><p> Bug reports should be submitted to <a href="http://www.linux-ha.org/ClusterResourceManager/BugReports" target="_top">http://www.linux-ha.org/ClusterResourceManager/BugReports</a>. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-architecture"></a>A.14.3. DRBD Architecture</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD Architecture. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-3-1">28.14.3.1: </a> Is an Active/Active option available for MySQL with DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-3-2">28.14.3.2: </a> What MySQL storage engines are supported with DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-3-3">28.14.3.3: </a> How long does a failover take? </p></li><li><p><a href="faqs.html#qandaitem-28-14-3-4">28.14.3.4: </a> How long does it take to resynchronize data after a failure? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-3-1"></a><span class="bold"><strong>28.14.3.1: </strong></span><span class="bold"><strong> Is an Active/Active option available for MySQL with DRBD? </strong></span></p><p> Currently, MySQL does not support Active/Active configurations using DRBD “<span class="quote">out of the box</span>”. </p><p><a name="qandaitem-28-14-3-2"></a><span class="bold"><strong>28.14.3.2: </strong></span><span class="bold"><strong> What MySQL storage engines are supported with DRBD? </strong></span></p><p> All of the MySQL transactional storage engines are supported by DRBD, including InnoDB and Falcon. For archived or read-only data, MyISAM or Archive can also be used. </p><p><a name="qandaitem-28-14-3-3"></a><span class="bold"><strong>28.14.3.3: </strong></span><span class="bold"><strong> How long does a failover take? </strong></span></p><p> Failover time is dependent on many things, some of which are configurable. After activating the passive host, MySQL will have to start and run a normal recovery process. If the InnoDB log files have been configured to a large size and there was heavy write traffic, this may take a reasonably long period of time. However, under normal circumstances, failover tends to take less than a minute. </p><p><a name="qandaitem-28-14-3-4"></a><span class="bold"><strong>28.14.3.4: </strong></span><span class="bold"><strong> How long does it take to resynchronize data after a failure? </strong></span></p><p> Resynchronization time depends on how long the two machines are out of communication and how much data was written during that period of time. Resynchronization time is a function of data to be synced, network speed and disk speed. DRBD maintains a bitmap of changed blocks on the primary machine, so only those blocks that have changed will need to be transferred. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-mysql-replication-scale"></a>A.14.4. DRBD and MySQL Replication</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about MySQL Replication Scale-out. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-4-1">28.14.4.1: </a> What is the difference between MySQL Cluster and DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-4-2">28.14.4.2: </a> What is the difference between MySQL Replication and DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-4-3">28.14.4.3: </a> How can I combine MySQL Replication scale-out with DRBD? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-4-1"></a><span class="bold"><strong>28.14.4.1: </strong></span><span class="bold"><strong> What is the difference between MySQL Cluster and DRBD? </strong></span></p><p> Both MySQL Cluster and DRBD replicate data synchronously. MySQL Cluster leverages a shared-nothing storage architecture in which the cluster can be architected beyond an Active/Passive configuration. DRBD operates at a much lower level within the “<span class="quote">stack</span>”, at the disk I/O level. For a comparison of various high availability features between these two options, please refer to <a href="ha-overview.html" title="Chapter 14. High Availability, Scalability, and DRBD">Chapter 14, <i>High Availability, Scalability, and DRBD</i></a>. </p><p><a name="qandaitem-28-14-4-2"></a><span class="bold"><strong>28.14.4.2: </strong></span><span class="bold"><strong> What is the difference between MySQL Replication and DRBD? </strong></span></p><p> MySQL Replication replicates data asynchronously while DRBD replicates data synchronously. For a comparison of various high availability features between these two options, please refer to the high availability comparison grid, <a href="ha-overview.html" title="Chapter 14. High Availability, Scalability, and DRBD">Chapter 14, <i>High Availability, Scalability, and DRBD</i></a>. </p><p><a name="qandaitem-28-14-4-3"></a><span class="bold"><strong>28.14.4.3: </strong></span><span class="bold"><strong> How can I combine MySQL Replication scale-out with DRBD? </strong></span></p><p> MySQL Replication is typically deployed in a Master to many Slaves configuration. In this configuration, having many Slaves provides read scalability. DRBD is used to provide high-availability for the Master MySQL Server in an Active/Passive configuration. This provides for automatic failover, safeguards against data loss, and automatically synchronizes the failed MySQL Master after a failover. </p><p> The most likely scenario in which MySQL Replication scale-out can be leveraged with DRBD is in the form of attaching replicated MySQL “<span class="quote">read-slaves</span>” off of the Active-Master MySQL Server, shown in <a href="faqs.html#active-master-mysql-server" title="Figure A.1. Active-Master MySQL server">Figure A.1, “Active-Master MySQL server”</a>. Since DRBD replicates an entire block device, master information such as the binary logs are also replicated. In this way, all of the slaves can attach to the Virtual IP Address managed by Linux Heartbeat. In the event of a failure, the asynchronous nature of MySQL Replication allows the slaves to continue with the new Active machine as their master with no intervention needed. </p><div class="figure"><a name="active-master-mysql-server"></a><p class="title"><b>Figure A.1. Active-Master MySQL server</b></p><div class="mediaobject"><img src="images/active-master-mysql-server.png" alt="Active-Master MySQL server"></div></div></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-file-systems"></a>A.14.5. DRBD and File Systems</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD and file systems. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-5-1">28.14.5.1: </a> Can XFS be used with DRBD? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-5-1"></a><span class="bold"><strong>28.14.5.1: </strong></span><span class="bold"><strong> Can XFS be used with DRBD? </strong></span></p><p> Yes. XFS uses dynamic block size, thus DRBD 0.7 or later is needed. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-lvm"></a>A.14.6. DRBD and LVM</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD and LVM. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-6-1">28.14.6.1: </a> Can I use DRBD on top of LVM? </p></li><li><p><a href="faqs.html#qandaitem-28-14-6-2">28.14.6.2: </a> Can I use LVM on top of DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-6-3">28.14.6.3: </a> Can I use DRBD on top of LVM while at the same time running LVM on top of that DRBD? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-6-1"></a><span class="bold"><strong>28.14.6.1: </strong></span><span class="bold"><strong> Can I use DRBD on top of LVM? </strong></span></p><p> Yes, DRBD supports on-line resizing. If you enlarge your logical volume that acts as a backing device for DRBD, you can enlarge DRBD itself too, and of course your file system if it supports resizing. </p><p><a name="qandaitem-28-14-6-2"></a><span class="bold"><strong>28.14.6.2: </strong></span><span class="bold"><strong> Can I use LVM on top of DRBD? </strong></span></p><p> Yes, you can use DRBD as a Physical Volume (PV) for LVM. Depending on the default LVM configuration shipped with your distribution, you may need to add the <code class="filename">/dev/drbd*</code> device files to the <code class="literal">filter</code> option in your <code class="filename">lvm.conf</code> so LVM scans your DRBDs for PV signatures. </p><p><a name="qandaitem-28-14-6-3"></a><span class="bold"><strong>28.14.6.3: </strong></span><span class="bold"><strong> Can I use DRBD on top of LVM while at the same time running LVM on top of that DRBD? </strong></span></p><p> This requires careful tuning of your LVM configuration to avoid duplicate PV scans, but yes, it is possible. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-virtualization"></a>A.14.7. DRBD and Virtualization</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD and virtualization. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-7-1">28.14.7.1: </a> Can I use DRBD with OpenVZ? </p></li><li><p><a href="faqs.html#qandaitem-28-14-7-2">28.14.7.2: </a> Can I use DRBD with Xen and/or KVM? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-7-1"></a><span class="bold"><strong>28.14.7.1: </strong></span><span class="bold"><strong> Can I use DRBD with OpenVZ? </strong></span></p><p> See <a href="http://wiki.openvz.org/HA_cluster_with_DRBD_and_Heartbeat" target="_top">http://wiki.openvz.org/HA_cluster_with_DRBD_and_Heartbeat</a>. </p><p><a name="qandaitem-28-14-7-2"></a><span class="bold"><strong>28.14.7.2: </strong></span><span class="bold"><strong> Can I use DRBD with Xen and/or KVM? </strong></span></p><p> Yes. If you are looking for professional consultancy or expert commercial support for Xen- or KVM-based virtualization clusters with DRBD, contact LINBIT (<a href="http://www.linbit.com" target="_top">http://www.linbit.com</a>). </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-security"></a>A.14.8. DRBD and Security</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD and security. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-8-1">28.14.8.1: </a> Can I encrypt/compress the exchanged data? </p></li><li><p><a href="faqs.html#qandaitem-28-14-8-2">28.14.8.2: </a> Does DRBD do mutual node authentication? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-8-1"></a><span class="bold"><strong>28.14.8.1: </strong></span><span class="bold"><strong> Can I encrypt/compress the exchanged data? </strong></span></p><p> Yes. But there is no option within DRBD to allow for this. You’ll need to leverage a VPN and the network layer should do the rest. </p><p><a name="qandaitem-28-14-8-2"></a><span class="bold"><strong>28.14.8.2: </strong></span><span class="bold"><strong> Does DRBD do mutual node authentication? </strong></span></p><p> Yes, starting with DRBD 8 shared-secret mutual node authentication is supported. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-system-requirements"></a>A.14.9. DRBD and System Requirements</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD and System Requirements. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-9-1">28.14.9.1: </a> What other packages besides DRBD are required? </p></li><li><p><a href="faqs.html#qandaitem-28-14-9-2">28.14.9.2: </a> How many machines are required to set up DRBD? </p></li><li><p><a href="faqs.html#qandaitem-28-14-9-3">28.14.9.3: </a> Does DRBD only run on Linux? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-9-1"></a><span class="bold"><strong>28.14.9.1: </strong></span><span class="bold"><strong> What other packages besides DRBD are required? </strong></span></p><p> When using pre-built binary packages, none except a matching kernel, plus packages for <code class="literal">glibc</code> and your favorite shell. When compiling DRBD from source additional prerequisite packages may be required. They include but are not limited to: </p><div class="itemizedlist"><ul type="disc"><li><p> glib-devel </p></li><li><p> openssl </p></li><li><p> devel </p></li><li><p> libgcrypt-devel </p></li><li><p> glib2-devel </p></li><li><p> pkgconfig </p></li><li><p> ncurses-devel </p></li><li><p> rpm-build </p></li><li><p> rpm-devel </p></li><li><p> redhat-rpm-config </p></li><li><p> gcc </p></li><li><p> gcc-c++ </p></li><li><p> bison </p></li><li><p> flex </p></li><li><p> gnutls-devel </p></li><li><p> lm_sensors-devel </p></li><li><p> net-snmp-devel </p></li><li><p> python-devel </p></li><li><p> bzip2-devel </p></li><li><p> libselinux-devel </p></li><li><p> perl-DBI </p></li><li><p> libnet </p></li></ul></div><p> Pre-built x86 and x86_64 packages for specific kernel versions are available with a support subscription from LINBIT. Please note that if the kernel is upgraded, DRBD must be as well. </p><p><a name="qandaitem-28-14-9-2"></a><span class="bold"><strong>28.14.9.2: </strong></span><span class="bold"><strong> How many machines are required to set up DRBD? </strong></span></p><p> Two machines are required to achieve the minimum degree of high availability. Although at any one given point in time one will be primary and one will be secondary, it is better to consider the machines as part of a mirrored pair without a “<span class="quote">natural</span>” primary machine. </p><p><a name="qandaitem-28-14-9-3"></a><span class="bold"><strong>28.14.9.3: </strong></span><span class="bold"><strong> Does DRBD only run on Linux? </strong></span></p><p> DRBD is a Linux Kernel Module, and can work with many popular Linux distributions. DRBD is currently not available for non-Linux operating systems. </p></div><div class="section" lang="en"><div class="titlepage"><div><div><h3 class="title"><a name="drbd-support-consulting"></a>A.14.10. DBRD and Support and Consulting</h3></div></div></div><p> In the following section, we provide answers to questions that are most frequently asked about DRBD and resources. </p><p><span class="bold"><strong>Questions</strong></span></p><div class="itemizedlist"><ul type="disc"><li><p><a href="faqs.html#qandaitem-28-14-10-1">28.14.10.1: </a> Does MySQL offer professional consulting to help with designing a DRBD system? </p></li><li><p><a href="faqs.html#qandaitem-28-14-10-2">28.14.10.2: </a> Does MySQL offer support for DRBD and Linux Heartbeat from MySQL? </p></li><li><p><a href="faqs.html#qandaitem-28-14-10-3">28.14.10.3: </a> Are pre-built binaries or RPMs available? </p></li><li><p><a href="faqs.html#qandaitem-28-14-10-4">28.14.10.4: </a> Does MySQL have documentation to help me with the installation and configuration of DRBD and Linux Heartbeat? </p></li><li><p><a href="faqs.html#qandaitem-28-14-10-5">28.14.10.5: </a> Is there a dedicated discussion forum for MySQL High-Availability? </p></li><li><p><a href="faqs.html#qandaitem-28-14-10-6">28.14.10.6: </a> Where can I get more information about MySQL for DRBD? </p></li></ul></div><p><span class="bold"><strong>Questions and Answers</strong></span></p><p><a name="qandaitem-28-14-10-1"></a><span class="bold"><strong>28.14.10.1: </strong></span><span class="bold"><strong> Does MySQL offer professional consulting to help with designing a DRBD system? </strong></span></p><p> Yes. MySQL offers consulting for the design, installation, configuration, and monitoring of high availability DRBD. For more information concerning a High Availability Jumpstart, please see: <a href="http://www.mysql.com/consulting/packaged/scaleout.html" target="_top">http://www.mysql.com/consulting/packaged/scaleout.html</a>. </p><p><a name="qandaitem-28-14-10-2"></a><span class="bold"><strong>28.14.10.2: </strong></span><span class="bold"><strong> Does MySQL offer support for DRBD and Linux Heartbeat from MySQL? </strong></span></p><p> Yes. Support for DRBD is available with an add-on subscription to MySQL Enterprise called “DRBD for MySQL”. For more information about support options for DRBD see: <a href="http://mysql.com/products/enterprise/features.html" target="_top">http://mysql.com/products/enterprise/features.html</a>. </p><p> For the list of supported Linux distributions, please see: <a href="http://www.mysql.com/support/supportedplatforms/enterprise.html" target="_top">http://www.mysql.com/support/supportedplatforms/enterprise.html</a>. </p><div class="note" style="margin-left: 0.5in; margin-right: 0.5in;"><h3 class="title">Note</h3><p> DRBD is only available on Linux. DRBD is not available on Windows, MacOS, Solaris, HPUX, AIX, FreeBSD, or other non-Linux platforms. </p></div><p><a name="qandaitem-28-14-10-3"></a><span class="bold"><strong>28.14.10.3: </strong></span><span class="bold"><strong> Are pre-built binaries or RPMs available? </strong></span></p><p> Yes. “<span class="quote">DRBD for MySQL</span>” is an add-on subscription to MySQL Enterprise, which provides pre-built binaries for DRBD. For more information see: <a href="http://mysql.com/products/enterprise/features.html" target="_top">http://mysql.com/products/enterprise/features.html</a>. </p><p><a name="qandaitem-28-14-10-4"></a><span class="bold"><strong>28.14.10.4: </strong></span><span class="bold"><strong> Does MySQL have documentation to help me with the installation and configuration of DRBD and Linux Heartbeat? </strong></span></p><p> Coming soon. </p><p><a name="qandaitem-28-14-10-5"></a><span class="bold"><strong>28.14.10.5: </strong></span><span class="bold"><strong> Is there a dedicated discussion forum for MySQL High-Availability? </strong></span></p><p> Yes, <a href="http://forums.mysql.com/list.php?144" target="_top">http://forums.mysql.com/list.php?144</a>. </p><p><a name="qandaitem-28-14-10-6"></a><span class="bold"><strong>28.14.10.6: </strong></span><span class="bold"><strong> Where can I get more information about MySQL for DRBD? </strong></span></p><p> For more information about MySQL for DRBD, including a technical white paper please see: <a href="http://www.mysql.com/products/enterprise/drbd.html" target="_top">DRBD for MySQL High Availability</a>. </p></div></div></div><div class="navfooter"><hr><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="extending-mysql.html">Prev</a> </td><td width="20%" align="center"> </td><td width="40%" align="right"> <a accesskey="n" href="error-handling.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Chapter 26. Extending MySQL </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> Appendix B. Errors, Error Codes, and Common Problems</td></tr></table></div></body></html>