MariaDB Server is an open-source project developed by MariaDB plc and stewarded by MariaDB Foundation. Anyone can participate in the development.

This section provides information to help you participate in making MariaDB Server and other MariaDB products a success.

If you're interested in contributing to the documentation, see this page.

How and where the community can report bugs and file feature requests, and how bug reports are processed for community users. MariaDB plc provides SLA for customer issues (see ).

To report documentation issues, please review .

There are many questions that are frequently asked about MariaDB, the organizations and the software. They have been grouped based on common topics. Click on a topic to find the question you may have. If you don't see it, please click on the button in the left margin labeled, "Ask a Question Here". We'll try to answer you question, quickly.

The 'MySQL' name is trademarked by Oracle, and they have chosen to keep that trademark to themselves. The name MySQL (just like the MyISAM storage engine) comes from Monty's first daughter My. The first part of 'MySQL' is pronounced like the English adjective, even if this doesn't match the correct pronunciation of the Finnish name.

MariaDB continues this tradition by being named after his younger daughter, Maria.

The name Maria was initially given to a storage engine. After MariaDB was started, to avoid confusion, it was renamed to Aria. The new name was decided as a result of a contest.

_{This page is licensed: CC BY-SA / Gnu FDL}

Hi all,

I have create a remote connection between two linux server running mariadb. One run centos stream 9 with 10.5.16-MariaDB and the second run centos 7 with 5.5.68-MariaDB. Through php script automated with cron I run queries and it work.

The problem is that often i get the error HY000/2002 connection timed out. The error occurs at random times either using cron or launching it manually. For example, the script work for 10 times and on the 11th time I receive the error (HY000/2002 connection timed out) which can occur one or more times consecutively and then return to work on the umpteenth execution. The number of times the script works or not is always different.

I can't find the cause of the error. Any suggestions?

Thanks

_{This page is licensed: CC BY-SA / Gnu FDL}

You can read about the features and improvements included in the current stable release of MariaDB on the MariaDB versus MySQL page.

The MariaDB Development Plans page links to our plans for future versions of MariaDB.

_{This page is licensed: CC BY-SA / Gnu FDL}

The MariaDB Development Plans page links to plans for future versions of MariaDB.

The release schedule for upcoming MariaDB releases can be found on the MariaDB Jira release page.

See also the page.

_{This page is licensed: CC BY-SA / Gnu FDL}

See MariaDB Security Policy.

See Also

_{This page is licensed: CC BY-SA / Gnu FDL}

See MariaDB Development for this.

_{This page is licensed: CC BY-SA / Gnu FDL}

All versions of MariaDB are based on MySQL 5.1 and greater, thus the table_cache option is deprecated in favor of table_open_cache. This is also documented at: .

For further reading, please refer to the MySQL manual: How MySQL Opens and Closes Tables.

Examples of use cases:

MariaDB [(none)]> SHOW GLOBAL STATUS LIKE 'opened_tables';
+---------------+--------+
| Variable_name | Value  |
+---------------+--------+
| Opened_tables | 354858 |
+---------------+--------+
1 row in set (0.00 sec)

MariaDB [(none)]> SELECT @@table_open_cache;
+--------------------+
| @@table_open_cache |
+--------------------+
|                400 |
+--------------------+
1 row in set (0.00 sec)

_{This page is licensed: CC BY-SA / Gnu FDL}

There are many ways to contribute to MariaDB. If you want to contribute to, or participate in the development of MariaDB, there are many ways to do so. You don't have to be a developer (but we always welcome more of those), you just have to have the willingness to help make MariaDB better. For example, we are looking for writers or translators of KB articles and people helping setting up MariaDB discussions groups.

Contributing to the MariaDB Project is the page that gathers together everything you need to get started.

The community category contains a lot of interesting links about how to participate.

You can also consider sponsoring a feature.

Welcome to the MariaDB community!

_{This page is licensed: CC BY-SA / Gnu FDL}

For reporting documentation bugs specifically, see .

MariaDB's bug and feature tracker is found at .

This page contains general guidelines for the community for reporting bugs and feature requests in MariaDB products. If you want to discuss a problem or a new feature with other MariaDB developers, you can find the email lists and forums .

Known Issues

First, check that the bug or feature request isn't already filed in the .

For the MariaDB issue database, use JIRA search to check if a report you are going to submit already exists. You are not expected to be a JIRA search guru, but please at least make some effort.

MariaDB is a , drop-in replacement of the MySQL® Database Server. It includes all major open source storage engines.

You can see a as well as a .

The source code for MariaDB is on . Binaries and packages are .

_{This page is licensed: CC BY-SA / Gnu FDL}

To provide a developed, stable, and DBMS that is, on the user level, broadly compatible with MySQL.

We strive for interoperability with both our own, and our upstream, communities.

Until , MariaDB was kept up to date with the latest MySQL release from the same branch. For example was up to date with MySQL 5.1.47, and so on.

We did a merge from the main MySQL branch for every new MySQL release or when there was some critical bugfix applied to the main branch.

Since , MariaDB includes backported features from MySQL as well as entirely new features not found anywhere else, but does not necessarily include all MySQL features.

We strive to keep our main trees as free from bugs as possible. It should be reasonably safe to pull from our trees at any time.

_{This page is licensed: CC BY-SA / Gnu FDL}

The code is hosted on . You can branch the latest code from there, and you can also push your own changes as a new branch that can be shared with others.

is done with standard Unix tools: CMake (or autotools for MariaDB versions below 5.5), Gnu Make, GCC (or other C/C++ compiler on some systems). On , CMake and Visual Studio are used.

The current state of the source with respect to build/test failures can be seen in .

For project management and bug tracking, we use .

The page has links to instructions on setting up a full development environment, if you are interested.

_{This page is licensed: CC BY-SA / Gnu FDL}

According to MySQL docs:

The maximum number of tables that can be referenced in a single join is 61. This also applies to the number of tables that can be referenced in the definition of a view.

Does the same maximum (61) applies to MariaDB, too?

If yes, are there plans for raising this limit?

_{This page is licensed: CC BY-SA / Gnu FDL}

• Q: What is MariaDB Community Server and its main purpose? A: MariaDB Community Server is a popular free and open-source relational database management system (RDBMS), originally forked from MySQL. It is actively developed and maintained by MariaDB plc and a global community of contributors. It's known for its strong performance, stability, and a rich feature set suitable for a wide range of applications.

• Q: Can MariaDB Community Server be used as a direct MySQL replacement? A: Yes, in many scenarios and for numerous versions, MariaDB Community Server is designed to be a compatible, drop-in replacement for MySQL, facilitating easy migration for users. However, as both MariaDB and MySQL evolve independently, some differences may exist in newer or more advanced features.

• Q: What are the key benefits and features of MariaDB Community Server? A: Key benefits of MariaDB Community Server include its true open-source nature (GPLv2 license), a broader selection of advanced storage engines (like Aria, MyRocks, and Spider, in addition to the standard InnoDB), continuous feature enhancements, performance improvements, and the support of a vibrant, active global community.

Here you can ask questions if MariaDB supports a certain feature that doesn't fit into any specific category.

Before asking a questions about a non existing feature, please first check the if there is already plans for that feature.

The page has the current stable and development releases.

See the page for plans for future versions of MariaDB.

_{This page is licensed: CC BY-SA / Gnu FDL}

is part of the Amazon Web Services, that allows one to setup and run a relational database in the cloud. It supports running MariaDB in addition to MySQL, PostgreSQL, and other common relational databases.

Gee-Hwan Chuang of Listia has written a guide titled . He recommends using .

_{This page is licensed: CC BY-SA / Gnu FDL}

What is the Intention of the MCA?

1. The original copyright holder and the MariaDB Foundation both get (shared) copyright for the code.

2. Both parties can use their copy independently of each other for any purpose (as if each copy was completely separate and unique).

What if I Don't Want to Submit Code to You Under the ?

If you don't like the for code that you want to contribute to , the other option is to submit your code under the license.

What is the difference between MCA and BSD-new?

The main difference is that MCA is a copyright assignment agreement; you assign the copyright to the MariaDB Foundation (but you retain it too, on your copy of the contribution), while BSD-new is a license; you allow the MariaDB Foundation to use your contribution under specific (albeit rather broad) terms.

When should I use MCA and when should I use BSD-new?

With MCA, the MariaDB Foundation becomes a copyright holder, and it can defend the GPL and pursue violators, protecting its and your rights. This is the same reason why the Free Software Foundation too.

On the other hand, BSD-new is a well known FSF-approved Open Source license.

If you contribute on behalf of your employer, your employer's legal department is most probably already familiar with the BSD-new license. It might be much easier and faster to get them to approve your contribution if you contribute under this license.

If you are an individual contributor (or if your employer is fine with MCA), it makes sense to contribute under MCA, as it will allow the MariaDB Foundation to defend its Open Source way when needed.

For most other practical purposes there is no difference between these two ways of contributing. You keep the copyright to your contribution. And the MariaDB Foundation will be able to use your contribution in MariaDB Server, relicense it (for example, as GPLv2 or GPLv3) or use separately as a patch.

What Will the MariaDB Foundation Do With the Code I Contribute?

The MariaDB Foundation agrees that when it dual licenses code, it will not restrict the way the third party licensee uses the licensed copy of the code nor restrict how they use their own code.

Where Did the Come From?

The is based on . We think it is one of the best contributor agreements around!

What's the Difference Between Sun's SCA and the ?

Sun's SCA was unclear in a couple of points which we have fixed in the :

• The patent grant has been made clear that it only affects the donated code, not any other code.

• We added a clarification how code is donated; "Each submission must explicitly be marked that it's donated under the MCA"

• We have made a promise in this FAQ how the MariaDB Foundation will Dual-license code. See Monty's blog post "" for a discussion about this.

What Happened to the Monty Program Contributor Agreement?

With the establishment of the MariaDB Foundation, Monty Program Ab transferred all existing signed contributor agreements to the Foundation.

Other Questions?

Question not answered? Subscribe to the and ask there or visit us on our or the #maria channel on.

_{This page is licensed: CC BY-SA / Gnu FDL}

Question: When I use the mysql client as in mysql -u root -p -h <host>, does the password-enter prompt popup before or after the client has established a connection with the server?

Answer: Before the client has established a connection to the server.

How do we derive the answer?

lovegood:~ byte$ mysql -uroot -p -hlocalhost
Enter password: 
Welcome to the MariaDB monitor.  Commands end with ; or \g.
Your MariaDB connection id is 1
Server version: 5.2.8-MariaDB Source distribution

This software comes with ABSOLUTELY NO WARRANTY. This is free software,
and you are welcome to modify and redistribute it under the GPL v2 license

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

MariaDB [(none)]> \q
Bye
lovegood:~ byte$ mysql.server stop
Shutting down MySQL
. SUCCESS! 
lovegood:~ byte$ mysql -uroot -p -hlocalhost
Enter password: 
ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/tmp/mysql.sock' (2)

_{This page is licensed: CC BY-SA / Gnu FDL}

Instructions on how to build a mysqld that contains all the information we need to fix problems you encounter. (A more detailed explanation can be found here.)

• Add the option to your /.my.cnf or /etc/my.cnf file under the [mysqld] tag.

• Get the latest MariaDB code from GitHub.

• Compile MariaDB with the -g compiler flag (Unix).

• Optionally: with more checking - will cause slowdown.

• Shut down your old mysqld server.

• Install the new compiled mysqld binary. Note that if you are compiling same version of MariaDB that you have already installed it's enough to just copy this one binary!

• Restart mysqld.

Compiling with -g should not cause any notable slowdown of the server.

You can of course also do make install, but the above way allows you to go back to your old binary if needed.

If you get any errors about a wrong number of error messages, you can fix that by copying the corresponding language file from sql/share over your old ones (this should be reasonably safe to do).

What to Do When You Get a Crash After Installing a Debug Binary

Now when you get a crash do the following:

• Create a README file that describes the problem. You can use the mysqlbug script to generate a template for this.

• Create a tar file containing the core, the mysqld binary and README. If possible, also add any database files that could help us repeat the problem!

• Send it to our secure ftp server:

• To be able to follow the progress, create a bug report in about this. This should be easy to do based on the information you have in your README file.

See Also

• .

_{This page is licensed: CC BY-SA / Gnu FDL}

• Q: What is mariadb-backup and its primary use? A: mariadb-backup is a command-line utility included with MariaDB Community Server. Its primary use is to perform physical "hot" backups of MariaDB database instances, especially those utilizing transactional storage engines like InnoDB and XtraDB. This tool is based on the well-regarded Percona XtraBackup technology.

• Q: What are the main advantages of using mariadb-backup for database backups? A: The main advantages of mariadb-backup are its ability to perform online (hot) backups with minimal locking. This means your MariaDB database remains largely available for both read and write operations during the backup process. For large databases, mariadb-backup is generally much faster than logical backup methods (like mariadb-dump) and, when used in conjunction with MariaDB's binary logs, it enables precise point-in-time recovery (PITR).

• Q: How does the mariadb-backup utility work to create backups? A: mariadb-backup works by copying the physical data files from the MariaDB data directory while the server is running. It continuously monitors for changes made to these files during the backup operation and records these modifications in a separate log file. In a subsequent "prepare" phase, these logged changes are applied to the copied data files to ensure they are brought to a transactionally consistent state, ready for restoration.

• Q: Is mariadb-backup a suitable backup solution for all MariaDB storage engines? A: mariadb-backup is primarily designed for, and works most effectively with, transactional storage engines such as InnoDB and XtraDB, for which it can guarantee consistency and perform online backups. While it might be able to copy data files for other storage engines like Aria or MyISAM, it cannot ensure the same level of transactional consistency or perform true online backups for them. For non-transactional engines, mariadb-dump or filesystem-level snapshots might be more appropriate choices.

• Q: Is the mariadb-backup tool free to use? A: Yes, mariadb-backup is an open-source tool that is included as a standard utility with MariaDB Community Server and is completely free to use.

When upgrading from MySQL 5.1 to you may encounter a process limit problem with your MariaDB installation. This is not MariaDB at fault. The issue is likely triggered with high connection counts to the database and possibly exacerbated by not using Thread pooling.

You might encounter the following error message if you are not out of available memory:

##SQLSTATE[HY000] [1135] Can't create a new thread (errno 11);##

You can consult the manual for a possible OS-dependent bug when trying to connect. There is a very good explanation and troubleshooting with a work around on the Percona MySQL Performance Blog

Review your connections and thread consumption usage prior to doing an upgrade to avoid this being a problem.

_{This page is licensed: CC BY-SA / Gnu FDL}

Instructions to narrow down problems on a replication replica.

Sometimes a event causes an error of some sort. A whole binary log file is sometimes impractical due to size or sensitivity reasons.

Step 1: Copy the binary log locally

This is just in case you don't quite extract the right information first. If the binlog expired off and you haven't got the right information, your bug report may not easily be reproducible.

Step 2: Create an extract header

Binary logs have a header portion. Without the header won't be able to read it. The header also contains valuable session information

We look at the binary log to see how big the header and session information is:

We see that the session information ends at 328 because of the last line, so we extract to that point.

We need to find out at what offset the entry at 129619 ends and it might be useful to extract some previous entries as well.

This page describes how community bug reports are processed among our products and explains what you need to notice while tracking bugs.

Commitments

MariaDB does not have any SLA or guaranteed reaction times on bugs in Jira. While we are taking bugs reported by the community very seriously, and aim to provide response and to handle issues as fast as possible, MariaDB does not have a dedicated bug verification team, this activity is performed on the best-effort basis.

To make sure your bug report will be confirmed and moved forward faster, please follow about creating bug reports.

Questions? See the page.

MariaDB Contributor Agreement (MCA)

These terms apply to your contribution of materials to a product or project owned or managed by us ('project'), and set out the intellectual property rights you grant to us (The MariaDB Foundation) in the contributed material. If this contribution is on behalf of a company, the term 'you' will also mean the company you identify below. If you agree to be bound by these terms, fill in the information requested below and provide your signature. Read this agreement carefully before signing.

Query with ORDER BY in a FROM subquery produces unordered result. Is this a bug? Below is an example of this:

returns a result set that is not necessarily ordered by field2. This is not a bug.

A "table" (and subquery in the FROM clause too) is - according to the SQL standard - an unordered set of rows. Rows in a table (or in a subquery in the FROM clause) do not come in any specific order. That's why the optimizer can ignore the clause that you have specified. In fact, the SQL standard does not even allow the ORDER BY clause to appear in this subquery (we allow it, because ORDER BY ... LIMIT ... changes the result, the set of rows, not only their order).

You need to treat the subquery in the FROM clause, as a set of rows in some unspecified and undefined order, and put the ORDER BY on the top-level

If you can't find help in the , you can also subscribe to the or to communicate with MariaDB users and developers.

There's an active help community on for programming with SQL and indexing questions, and for non-programming questions for example tuning/backups/architecture/migration/replication etc. Use the mariadb tag so questions are searchable by those that can help you.

If you have a question about a feature that is not properly documented or something that is not working as expected, go to the corresponding and ask your question there.

You can report and check on bugs which apply to MariaDB in .

offers .

_{This page is licensed: CC BY-SA / Gnu FDL}

Questions asked by the public about MariaDB and related software. Note that bugs, problems and features requests should be submitted to the MariaDB bug database so that they can be process properly by the MariaDB developer and QA teams.

Profiling the memory usage can be useful for finding out why a program appears to use more memory than it should. It is especially helpful for analyzing OOM situations or other cases where the memory grows linearly and causes problems.

To profile the memory usage of a program, there are multiple options. The following sections describe the methods that are available.

If a problem in memory usage is identified and it appears to be due to a bug, please open a new bug report on the MariaDB Jira under the correct project and include the relevant memory profiling output in it. Refer to How to Write a Good Bug Report for more details.

Known issues that can cause extended memory usage

Transparent huge pages (THP)

Transparent huge pages (THP), which is enabled by default in many newer Linux distributions, can cause out-of-memory-issues for MariaDB as THP is not suitable for databases. This is described at .

MariaDB Community Server 10.6.17, MariaDB Enterprise Server 10.6.16-11 and all other MariaDB server releases after these have THP disabled.

System malloc is not good if there are a lot of allocations of different size.

If and do not show an increase in memory, but the process still increases in size, then a likely problem is the system memory allocation library (malloc). Replacing malloc with should fix the issue in this case.

Profiling with the MariaDB server

Recent MariaDB versions have a global variable that shows how much memory the MariaDB server has allocated. By monitoring this variable one can find out if if the MariaDB allocated memory grows.

One can also check memory usage per user with the :

This shows the current memory used per connection and the maximum memory they have used since the user connected.

The can also be used to find out who is allocated memory and for what.

Note that one can also set the variable to restrict a user's memory usage.

BPF Compiler Collection (bcc)

The toolkit comes with the memleak program that traces outstanding memory allocations. This is a very convenient way of debugging high memory usage as it'll immediately show where the memory is allocated at.

By default the tool will print output once every five seconds with the stacktraces that have the most open allocations. Ctrl+C can be used to interrupt the collection of the traces.

The profiling interval and the profiling duration can be passed as arguments to memleak. The first argument is how often a sample is taken and the second argument is how long to sample for. To help analyze excessive memory usage, collect the output of the memleak program for at least 60 seconds. The longer the profiling can be left on, the more accurate the information will be.

The overhead of the profiling can be large enough that it affects production workloads negatively. To reduce the overhead, the sampling frequency of memory allocations can be lowered using the --sample-rate option:

For example, -s 10 will sample only 10% of memory allocations which may miss out memory leaks from individual allocations but the longer the system is left running, the more likely it is that a leaking memory allocation is sampled. This means that even with a lower sampling rate, the source of the memory leak will eventually be found.

RHEL, CentOS, Rocky Linux and Fedora

On RHEL based systems, the package is named bcc-tools. After installing it, use the following command to profile the memory usage 5 times per second over a window of 60 seconds:

Ubuntu and Debian

On Ubuntu/Debian the package is named bpfcc-tools. After installing it, use the following command to profile the memory usage 5 times per second over a window of 60 seconds:

Jemalloc Heap Profiling

Jemalloc is an alternative to the default glibc memory allocator. It is capable of analyzing the heap memory usage of a process which allows it to be used to detect all sorts of memory usage problems with a lower overhead compared to tools like Valgrind. Unlike the ASAN and LSAN sanitizers, it is capable of detecting cases where memory doesn't actually leak but keeps growing with no upper limit (e.g. items get appended to a list but are never removed).

Ubuntu and Debian

To enable jemalloc, the packages for it must be first installed from the system repositories. Ubuntu 20.04 requires the following packages to be installed for jemalloc profiling:

RHEL, CentOS, Rocky Linux and Fedora

The version of jemalloc that is available in most Red Hat repositories is not compiled with memory profiling support enabled. For RHEL based distributions, the only option is to .

Configuring Jemalloc for Heap Profiling

Once installed, edit the systemd service file with systemctl edit mariadb.service and add the following lines into it. The path to the libjemalloc.so file is OS-specific so make sure it points to the correct file. The example here is for Ubuntu and Debian environments.

Then create the directory for the profile files:

And finally restart MariaDB with systemctl restart mariadb.service.

The directory in /var/lib/mysql/jeprof/ will start to be filled by versioned files with a .heap suffix. Every time the virtual memory usage reaches a new high, a file will be created. Initially, the files will be created very often but eventually the pace will slow down. Once the problematic memory usage has been identified, the latest .heap file can be analyzed with the jeprof program.

The simplest method is to generate a text report with the following command.

A better way to look at the generated heap profile is with the PDF output. However, this requires the installation of extra packages (apt -y install graphviz ghostscript gv). To generate the PDF report of the latest heap dump, run the following command:

The generated heap-report.pdf will contain a breakdown of the memory usage.

Note that the report generation with the jeprof program must be done on the same system where the profiling was done. If done elsewhere, the binaries do not necessarily match and can cause the report generation to fail.

Tcmalloc Heap Profiling

Similarly to the jemalloc memory allocator, the memory allocator comes with a leak checker and heap profiler.

Installation

RHEL, CentOS and Rocky Linux

On RHEL based systems, the gperftools package is in the EPEL repositories. These must be first enabled by installing the epel-release package.

After this, the gperftools package can be installed.

Ubuntu 20.04

Service file configuration

Once tcmalloc is installed, edit the systemd service file with systemctl edit mariadb.service and add the following lines into it.

Note: Make sure to use the correct path and library name to the tcmalloc library in LD_PRELOAD. The following example uses the Debian location of the library. The file is usually located in /usr/lib64/libtcmalloc_and_profiler.so.4 on RHEL systems. The version number of the library can also change which might require other adjustments to the library path.

Then create the directory for the profile files:

And finally restart MariaDB with systemctl restart mariadb.service.

Configuring Heap Dump Frequency

The heap profiling is configured using environment variables. The details can be found in the Modifying Runtime Behavior section of the gperftools documentation:

By default, tcmalloc dumps the heap profile every time 1GiB of memory has been allocated (HEAP_PROFILE_ALLOCATION_INTERVAL) or whenever the high-water memory usage mark increases by 100MiB (HEAP_PROFILE_INUSE_INTERVAL). If there's no activity, no memory dumps will be generated.

To trigger a memory dump based on a time interval, set the HEAP_PROFILE_TIME_INTERVAL environment variable to the number of seconds between each dump. For example, with Environment=HEAP_PROFILE_TIME_INTERVAL=3600 there will be one heap dump per hour.

Report generation

Depending on which OS you are using, the report generation program is named either pprof (RHEL) or google-pprof (Debian/Ubuntu).

It is important to pick the latest .heap file to analyze. The following command generates the heap-report.pdf from the latest heap dump. The file will show the breakdown of the memory usage.

See also

_{This page is licensed: CC BY-SA / Gnu FDL}

In this article we will explain how the MariaDB and MySQL server and client library licenses affect other applications and what steps you can take to comply with the GPL or avoid having to change your applications to GPL.

In the following text we talk about MariaDB. The guidelines also apply to MySQL in general, at least up to 5.1.55 which was the latest stable release at the time of writing.

Licenses Used by MariaDB

MariaDB is distributed under the GPL license, version 2.

The MariaDB client libraries for , and are distributed under the LGPL license, version 2.1 or later. The LGPL license allows you to distribute these MariaDB client libraries freely with any application.

The MariaDB client library included with the MariaDB server is also GPL version 2, but has a that allows you to combine it with most other open source software, without conflicting with their license, even if that license is incompatible with the GPL. We do however recommend you to use the new for any non-GPL application.

Internal Usage is Free

The GPL license only affects code that you distribute to other parties.

Internal usage within an organization is totally free and not subject to any conditions. There is no such thing as 'internal distribution' that would restrict the usage of your code by requiring it to be GPLed.

Connecting to a remote service that runs MariaDB (or any other GPL software) in the background is also free.

For internal programs for which you own all the copyright(s), there is essentially no risk in using GPL software. The argument you can use in your defense is that if the software became GPL as part of the distribution, you as the copyright holder could immediately revert your part back to its original copyright. No one has the right to require you to reveal or redistribute your code to the outside of your organization even if you would have distributed it internally linked with GPL software!

If your lawyers are concerned about distributions of software linked with GPL libraries between different legal entities within your organization, you can solve this by distributing your components and the GPL software separately, and have your other entity combining them. You can also switch to use the new .

Distributing an Application with a MariaDB Connector/Client

This section is for those that want to distribute the MariaDB client library code, but not the server, with their applications.

Free Software/Open-Source Applications

If your application is Free software/open source and uses one of the licenses listed in the , the GPL in the client library does not affect your application.

In other cases we recommend you to use the new .

Using a Connector That is not GPL

If you are using a connector that is not GPL, you are only bound by the license of that connector. Some examples are:

• .

The above have licenses that allow you to use them freely, without you being bound by the GPL.

Using a Database-Source-Independent Framework

If you are using a framework that allows you to connect dynamically to different RDBMS systems, any GPL licensed module loaded by the framework will not affect the application. Such frameworks are

• ODBC (Open Database Connectivity)

• JDBC (Java Database connectivity)

• Perl

The reason the GPL in the MySQL client library would not affect your application in this case is that the GPL client is supporting a standard interface and is thus merely an optional component among many. Your application could just as easily use the framework to connect to a RDBMS other than MariaDB or MySQL.

Any software can be connected to the GPL v2 licensed MySQL Connector/ODBC, without the need for that software to be GPLed. This is because there is a piece of general management software, the ODBC manager, between the GPLed MySQL Connector/ODBC and your software. If any logic would require the software which interfaces with MySQL Connector/ODBC to be GPL, then that would apply also to the ODBC manager itself. Yet, the ODBC manager is not GPL, neither on Windows nor on Linux. By consequence, no one would be allowed to use MySQL ODBC driver for anything.

Using the MariaDB Client Library for C

If your application is using a license that is not covered by the , you should use the new .

The LGPL license allows you to distribute these MariaDB client library freely with any application. If you modify the client library, you need to publish the new source code.

Distributing a Proprietary Application With the MariaDB / MySQL Server

When you are distributing your application together with MariaDB or MySQL, you are bound by the GPL if some of the following statements apply:

• You are using GPL code from MySQL linked directly to your application. (Like the MySQL GPL client library).

• Your application requires the MariaDB server to work and without the MariaDB server it doesn't start or it has very limited functionality.

The problem with the client library can be avoided by using one of the solutions mentioned before on this page.

If your application works with many databases, either natively or by using one of the , then you can freely distribute the MariaDB server with your application without being affected by the GPL. The reason for this is that MariaDB would only be an optional, independent component in your software distribution and section 2 of the GPL explicitly allows this:

You also have the option to buy licenses for MySQL from Oracle to get MySQL under other copyright terms. If you would like to later be able to use MariaDB instead of MySQL, please ensure that your license agreement allows you to make changes to the MySQL code! (This is something that you should ensure in all cases as otherwise you may run into bugs that Oracle will not fix, you are not allowed to fix and could make MySQL software unusable for you!)

The rights to use the MariaDB code changes in your application can be requested.

Legal Notice

The text above is written by Michael "Monty" Widenius, who is not a lawyer and you should not regard any statements of the above as 'ultimate truth' in all scenarios. On the other hand, it was David and Monty who together decided to make MySQL GPL and also decided and openly declared the intentions behind this license change, so there is some merit to information in this article.

If you want a second opinion of how GPL works in this case, you can contact or about this. Neither part has had anything to do with this KB article but they are the default authorities to turn to when you want to know more about the GPL or LGPL.

_{This page is licensed: CC BY-SA / Gnu FDL}