MariaDB server is a community developed fork of MySQL server. Started by core members of the original MySQL team, MariaDB actively works with outside developers to deliver the most featureful, stable, and sanely licensed open SQL server in the industry.
Find a file
2013-10-07 10:40:40 -04:00
buildbot add buildbot compiler suppressions file 2013-07-10 14:35:59 -04:00
buildheader tokutek/ft-index#80 tokutek/ft-engine#94 impose an upper bound on loader memory reservations 2013-10-07 08:35:52 -04:00
cmake closes #5485 #5368 clean up cmake and get darwin portability: 2013-04-17 00:01:07 -04:00
cmake_modules refs #64 fix clang on linux problems 2013-09-17 11:30:40 -04:00
examples tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
ft refs #78 for minicrons with a period of longer than one second (checkpoints), 2013-10-07 10:17:03 -04:00
include tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
locktree tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
portability tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
scripts updated drd test script to use the drd build type Tokutek/ft-engine#48 2013-07-24 13:49:19 -04:00
src refs #82, fix CMakeLists.txt 2013-10-07 10:40:40 -04:00
third_party/xz-4.999.9beta refs #5138 merge 49cfc8d... from xz-utils. fixes incorrect usage of restrict keyword 2013-04-17 00:02:30 -04:00
toku_include tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
util tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
utils tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
.gitignore Update .gitignore for tags and jemalloc symlink 2013-05-30 16:04:17 -07:00
bash.suppressions [t:4814] merging cmake work to mainline 2013-04-17 00:00:25 -04:00
CMakeLists.txt remove c++ api and db-benchmark-test 2013-05-02 11:55:13 -04:00
COPYING refs #6566 add COPYING file to fractal tree 2013-04-17 00:01:36 -04:00
CTestConfig.cmake refs #4871 use lex1:8080 to submit results to, lets us use /etc/hosts to submit from other machines 2013-04-17 00:01:14 -04:00
CTestCustom.cmake #17 fix a memory leak 2013-06-24 17:06:15 -04:00
DartConfig.cmake closes #5784 don't run valgrind on tdb-recover because it's statically linked and we can't change --soname-synonyms for just one child 2013-04-17 00:01:24 -04:00
README-TOKUDB tokutek/ft-index#76 add US Patent 8,489,638 2013-10-04 16:49:53 -04:00
README.md fixed required cmake version #65 2013-09-12 09:00:47 -04:00

TokuKV

TokuKV is a high-performance, transactional key-value store, used in the TokuDB storage engine for MySQL and MariaDB.

TokuKV is provided as a shared library with an interface similar to Berkeley DB.

To build the full MySQL product, see the instructions for ft-engine. This document covers TokuKV only.

Building

TokuKV is built using CMake >= 2.8.9. Out-of-source builds are recommended. You need a C++11 compiler, though only GCC >= 4.7 and Apple's Clang are tested. You also need zlib and valgrind development packages (yum install valgrind-devel zlib-devel or apt-get install valgrind zlib1g-dev).

You will also need the source code for jemalloc, checked out in third_party/.

git clone git://github.com/Tokutek/ft-index.git ft-index
cd ft-index
git clone git://github.com/Tokutek/jemalloc.git third_party/jemalloc
mkdir build
cd build
CC=gcc47 CXX=g++47 cmake \
    -D CMAKE_BUILD_TYPE=Debug \
    -D USE_BDB=OFF \
    -D BUILD_TESTING=OFF \
    -D CMAKE_INSTALL_PREFIX=../prefix/ \
    ..
cmake --build . --target install

This will build libtokudb.so and libtokuportability.so and install it, some header files, and some examples to ft-index/prefix/. It will also build jemalloc and install it alongside these libraries, you should link to that if you are planning to run benchmarks or in production.

Platforms

TokuKV is supported on 64-bit Centos, should work on other 64-bit linux distributions, and may work on OSX 10.8 and FreeBSD. TokuKV is not supported on 32-bit systems.

Transparent hugepages is a feature in newer linux kernel versions that causes problems for the memory usage tracking calculations in TokuKV and can lead to memory overcommit. If you have this feature enabled, TokuKV will not start, and you should turn it off. If you want to run with transparent hugepages on, you can set an environment variable TOKU_HUGE_PAGES_OK=1, but only do this for testing, and only with a small cache size.

Examples

There are some sample programs that can use either TokuKV or Berkeley DB in the examples/ directory. Follow the above instructions to build and install TokuKV, and then look in the installed examples/ directory for instructions on building and running them.

Testing

TokuKV uses CTest for testing. The CDash testing dashboard is not currently public, but you can run the tests without submitting them.

There are some large data files not stored in the git repository, that will be made available soon. For now, the tests that use these files will not run.

Many of the tests are linked with both TokuKV and Berkeley DB, as a sanity check on the tests themselves. To build these tests, you will need Berkeley DB and its header files installed. If you do not have Berkeley DB installed, just don't pass USE_BDB=ON.

In the build directory from above:

cmake -D BUILD_TESTING=ON [-D USE_BDB=ON] ..
ctest -D ExperimentalStart \
      -D ExperimentalConfigure \
      -D ExperimentalBuild \
      -D ExperimentalTest

Contributing

Please report bugs in TokuKV here on github.

We have two publicly accessible mailing lists:

We are also available on IRC on freenode.net, in the #tokutek channel.

License

TokuKV is available under the GPL version 2, with slight modifications. See README-TOKUDB.