Skip to content

Latest commit

 

History

History
175 lines (121 loc) · 5.59 KB

DEVELOPMENT.md

File metadata and controls

175 lines (121 loc) · 5.59 KB

Development

This document will help to setup your development environment and running tests. If you encounter a problem, please file an issue.

Table of contents

Building Milvus with Docker

Official releases are built using Docker containers. To build Milvus using Docker please follow these instructions.

Building Milvus on a local OS/shell environment

The details below outline the hardware and software requirements for building on Linux.

Hardware Requirements

Milvus is written in Go and C++, compiling it can use a lot of resources. We recommend the following for any physical or virtual machine being used for building Milvus.

- 8GB of RAM
- 50GB of free disk space

Installing Required Software

In fact, all Linux distributions is available to develop Milvus. The following only contains commands on Ubuntu, because we mainly use it. If you develop Milvus on other distributions, you are welcome to improve this document.

Dependencies

  • Debian/Ubuntu
sudo apt update
sudo apt install -y build-essential ccache gfortran \
    libssl-dev zlib1g-dev python3-dev libcurl4-openssl-dev libtbb-dev\
    libboost-regex-dev libboost-program-options-dev libboost-system-dev \
    libboost-filesystem-dev libboost-serialization-dev libboost-python-dev
  • CentOS
sudo yum install -y epel-release centos-release-scl-rh && \
sudo yum install -y git make automake openssl-devel zlib-devel \
      libcurl-devel python3-devel \
      devtoolset-7-gcc devtoolset-7-gcc-c++ devtoolset-7-gcc-gfortran \
      llvm-toolset-7.0-clang llvm-toolset-7.0-clang-tools-extra \
      ccache lcov

echo "source scl_source enable devtoolset-7" | sudo tee -a /etc/profile.d/devtoolset-7.sh
echo "source scl_source enable llvm-toolset-7.0" | sudo tee -a /etc/profile.d/llvm-toolset-7.sh
echo "export CLANG_TOOLS_PATH=/opt/rh/llvm-toolset-7.0/root/usr/bin" | sudo tee -a /etc/profile.d/llvm-toolset-7.sh
source "/etc/profile.d/llvm-toolset-7.sh"

# Install tbb
git clone https://github.com/wjakob/tbb.git && \
cd tbb/build && \
cmake .. && make -j && \
sudo make install && \
cd ../../ && rm -rf tbb/

# Install boost
wget -q https://boostorg.jfrog.io/artifactory/main/release/1.65.1/source/boost_1_65_1.tar.gz && \
    tar zxf boost_1_65_1.tar.gz && cd boost_1_65_1 && \
    ./bootstrap.sh --prefix=/usr/local --with-toolset=gcc --without-libraries=python && \
    sudo ./b2 -j2 --prefix=/usr/local --without-python toolset=gcc install && \
    cd ../ && rm -rf ./boost_1_65_1*

Once you have finished, confirm that gcc and make are installed:

gcc --version
make --version

CMake

The algorithm library of Milvus, Knowhere is written in c++. CMake is required in the Milvus compilation. If you don't have it, please follow the instructions in the Installing CMake.

Confirm that cmake is available:

cmake --version

Go

Milvus is written in Go. If you don't have a Go development environment, please follow the instructions in the Go Getting Started guide.

Confirm that your GOPATH and GOBIN environment variables are correctly set as detailed in How to Write Go Code before proceeding.

go version

Docker & Docker Compose

Milvus depends on Etcd, Pulsar and minIO. Using Docker Compose to manage these is an easy way in a local development. To install Docker and Docker Compose in your development environment, follow the instructions from the Docker website below:

Building Milvus

To build the Milvus project, run the following command:

make all

If this command succeed, you will now have an executable at bin/milvus off of your Milvus project directory.

A Quick Start for Testing Milvus

Presubmission Verification

Presubmission verification provides a battery of checks and tests to give your pull request the best chance of being accepted. Developers need to run as many verification tests as possible locally.

To run all presubmission verification tests, use this command:

make verifiers

Unit Tests

Pull requests need to pass all unit tests. To run every unit test, use this command:

make unittest

E2E Tests

Milvus uses Python SDK to write test cases to verify the correctness of Milvus functions. Before run E2E tests, you need a running Milvus:

cd deployments/docker/dev
docker-compose up -d
cd ../../../
# Running Milvus
./scripts/start_standalone.sh

# or

./scripts/start_cluster.sh

To run E2E tests, use these command:

cd tests20/python_client
pip install -r requirements.txt
pytest --tags=L0 --workers 4

GitHub Flow

To check out code to work on, please refer to the GitHub Flow.