milvus/cpp
wxyu f48ffc58a1 MS-88 add support for arm arch
Former-commit-id: db52dc345c907874d4fde1c0f048bab3cdc207f5
2019-06-17 15:32:25 +08:00
..
cmake solve conflicts 2019-06-17 14:26:46 +08:00
conf rename config item 2019-06-17 10:00:05 +08:00
scripts remove old names from source code 2019-06-14 17:29:29 +08:00
src MS-88 add support for arm arch 2019-06-17 15:32:25 +08:00
thirdparty solve conflicts 2019-06-17 14:26:46 +08:00
unittest remove old names from source code 2019-06-14 17:29:29 +08:00
.gitignore remove old names from source code 2019-06-14 17:29:29 +08:00
build.sh remove old names from source code 2019-06-14 17:29:29 +08:00
CHANGELOG.md MS-88 add support for arm arch 2019-06-17 15:32:25 +08:00
CMakeLists.txt remove old names from source code 2019-06-14 17:29:29 +08:00
LICENSE.txt Add LICENSE.txt and RELEASE.md 2019-04-14 20:12:34 +08:00
README.md remove old names from source code 2019-06-14 17:29:29 +08:00
RELEASE.md Add LICENSE.txt and RELEASE.md 2019-04-14 20:12:34 +08:00
start_server.sh remove old names from source code 2019-06-14 17:29:29 +08:00
stop_server.sh remove old names from source code 2019-06-14 17:29:29 +08:00
version.h.macro remove old names from source code 2019-06-14 17:29:29 +08:00

Compilation

Step 1: install necessery tools

centos7 : 
    yum install gfortran flex bison
    
ubuntu16.04 : 
    sudo apt-get install gfortran flex bison

Step 2: build(output to cmake_build folder)

cmake_build/src/milvus_server is the server

cmake_build/src/libmilvus_engine.a is the static library

cd [sourcecode path]/cpp
./build.sh -t Debug
./build.sh -t Release
./build.sh -g # Build GPU version

If you encounter the following error when building: protocol https not supported or disabled in libcurl

  1. Install libcurl4-openssl-dev

  2. Install cmake 3.14:

    ./bootstrap --system-curl 
    make 
    sudo make install
    

To build unittest:

./build.sh -u
or
./build.sh --unittest

Launch server

Set config in cpp/conf/server_config.yaml

Then launch server with config: cd [build output path] start_server.sh stop_server.sh

Launch test_client(only for debug)

If you want to test remote api, you can build test_client. test_client use same config file with server:

cd [build output path]/test_client
test_client -c [sourcecode path]/cpp/conf/server_config.yaml

License Generate

Use get_sys_info to get system info file.

./get_sys_info                  # system.info will be generated

Use license_generator to generate license file.

./license_generator -s system.info -l system.license -b 2019-05-15 -e 2019-08-14                 

Copy the license file to path assigned by license_path in server config file.