If you use Windows, you need to create a virtual machine with Ubuntu. To start working with a virtual machine please install VirtualBox. You can download Ubuntu from the website: https://www.ubuntu.com/\#download. Please create a virtual machine from the downloaded image (you should reserve at least 4GB of RAM for it). To run a command-line terminal in Ubuntu, please locate a program containing the word “terminal” in its name (gnome-terminal, konsole etc.) or just press Ctrl+Alt+T.
You probably already have one, but if you don’t, please register at https://github.com. In case you do not have SSH keys, you should generate them and then upload them on GitHub. It is required for sending over your patches. It is also possible to use the same SSH keys that you use with any other SSH servers - probably you already have those.
Create a fork of ClickHouse repository. To do that please click on the “fork” button in the upper right corner at https://github.com/ClickHouse/ClickHouse. It will fork your own copy of ClickHouse/ClickHouse to your account.
The development process consists of first committing the intended changes into your fork of ClickHouse and then creating a “pull request” for these changes to be accepted into the main repository (ClickHouse/ClickHouse).
Next, you need to download the source files onto your working machine. This is called “to clone a repository” because it creates a local copy of the repository on your working machine.
Please note that ClickHouse repository uses `submodules`. That is what the references to additional repositories are called (i.e.external libraries on which the project depends). It means that when cloning the repository you need to specify the `--recursive` flag as in the example above. If the repository has been cloned without submodules, to download them you need to run the following:
It generally means that the SSH keys for connecting to GitHub are missing. These keys are normally located in `~/.ssh`. For SSH keys to be accepted you need to upload them in the settings section of GitHub UI.
You can also clone the repository via https protocol:
This, however, will not let you send your changes to the server. You can still use it temporarily and add the SSH keys later replacing the remote address of the repository with `git remote` command.
Next, check the version of CMake: `cmake --version`. If it is below 3.3, you should install a newer version from the website: https://cmake.org/download/.
ClickHouse uses several external libraries for building. All of them do not need to be installed separately as they are built together with ClickHouse from the sources located in the submodules. You can check the list in `contrib`.
Compilers GCC starting from version 9 and Clang version 8 or above are supported for building ClickHouse.
Official Yandex builds currently use GCC because it generates machine code of slightly better performance (yielding a difference of up to several percent according to our benchmarks). And Clang is more convenient for development usually. Though, our continuous integration (CI) platform runs checks for about a dozen of build combinations.
To install GCC on Ubuntu run: `sudo apt install gcc g++`
Check the version of gcc: `gcc --version`. If it is below 9, then follow the instruction here: https://clickhouse.tech/docs/en/development/build/\#install-gcc-9.
Now that you are ready to build ClickHouse we recommend you to create a separate directory `build` inside `ClickHouse` that will contain all of the build artefacts:
While inside the `build` directory, configure your build by running CMake. Before the first run, you need to define environment variables that specify compiler (version 9 gcc compiler in this example).
For a faster build, you can resort to the `debug` build type - a build with no optimizations. For that supply the following parameter `-D CMAKE_BUILD_TYPE=Debug`:
If you get the message: `ninja: error: loading 'build.ninja': No such file or directory`, it means that generating a build configuration has failed and you need to inspect the message above.
While building messages about protobuf files in libhdfs2 library like `libprotobuf WARNING` may show up. They affect nothing and are safe to be ignored.
In this case, ClickHouse will use config files located in the current directory. You can run `clickhouse server` from any directory specifying the path to a config file as a command-line parameter `--config-file`.
You can replace the production version of ClickHouse binary installed in your system with your custom-built ClickHouse binary. To do that install ClickHouse on your machine following the instructions from the official website. Next, run the following:
If you do not know which IDE to use, we recommend that you use CLion. CLion is commercial software, but it offers 30 days free trial period. It is also free of charge for students. CLion can be used both on Linux and on Mac OS X.
KDevelop and QTCreator are other great alternatives of an IDE for developing ClickHouse. KDevelop comes in as a very handy IDE although unstable. If KDevelop crashes after a while upon opening project, you should click “Stop All” button as soon as it has opened the list of project’s files. After doing so KDevelop should be fine to work with.
Just in case, it is worth mentioning that CLion creates `build` path on its own, it also on its own selects `debug` for build type, for configuration it uses a version of CMake that is defined in CLion and not the one installed by you, and finally, CLion will use `make` to run build tasks instead of `ninja`. This is normal behaviour, just keep that in mind to avoid confusion.
Developing ClickHouse often requires loading realistic datasets. It is particularly important for performance testing. We have a specially prepared set of anonymized data from Yandex.Metrica. It requires additionally some 3GB of free disk space. Note that this data is not required to accomplish most of the development tasks.
Navigate to your fork repository in GitHub’s UI. If you have been developing in a branch, you need to select that branch. There will be a “Pull request” button located on the screen. In essence, this means “create a request for accepting my changes into the main repository”.
A pull request can be created even if the work is not completed yet. In this case please put the word “WIP” (work in progress) at the beginning of the title, it can be changed later. This is useful for cooperative reviewing and discussion of changes as well as for running all of the available tests. It is important that you provide a brief description of your changes, it will later be used for generating release changelogs.
Testing will commence as soon as Yandex employees label your PR with a tag “can be tested”. The results of some first checks (e.g.code style) will come in within several minutes. Build check results will arrive within half an hour. And the main set of tests will report itself within an hour.
The system will prepare ClickHouse binary builds for your pull request individually. To retrieve these builds click the “Details” link next to “ClickHouse build check” entry in the list of checks. There you will find direct links to the built .deb packages of ClickHouse which you can deploy even on your production servers (if you have no fear).
Most probably some of the builds will fail at first times. This is due to the fact that we check builds both with gcc as well as with clang, with almost all of existing warnings (always with the `-Werror` flag) enabled for clang. On that same page, you can find all of the build logs so that you do not have to build ClickHouse in all of the possible ways.
