- vearch base compile environment image address: https://hub.docker.com/r/vearch/vearch-dev-env/tags
- vearch deploy image address: https://hub.docker.com/r/vearch/vearch/tags
- If deploy a docker start vearch, master, ps, router start together
cp vearch/config/config.toml .
nohup docker run -p 8817:8817 -p 9001:9001 -v $PWD/config.toml:/vearch/config.toml vearch/vearch:latest all &
- If distributed deploy, modify vearch/config/config.toml and start separately.
- Modify vearch/config/config.toml, refer the step 'Local Model'
- Start separately image, modify step i 'all' to 'master' and 'ps' and 'router', master image must first start
- take vearch-dev-env:latest as an example
- docker pull vearch/vearch-dev-env:latest
- sh vearch/cloud/complile.sh
- sh build.sh
- reference "Use vearch image deploy" step 3
- build compile base environment image
- go to $vearch/cloud/env dir
- run
docker build -t vearch/vearch-dev-env:latest .
you will got a image named vearch-dev-env
- compile vearch
- go to $vearch/cloud dir
- run ./compile.sh you will compile Vearch in $vearch/build/bin , $vearch/build/lib
- make vearch image
- go to $vearch/cloud dir
- run ./build.sh you will got a image named vearch good luck
- how to use it
- you can use docker run -it -v config.toml:/vearch/config.toml vearch all to start vearch by local model the last param has four type[ps, router ,master, all] all means tree type to start
- One-click build vearch image
- go to $vearch/cloud dir
- you can run ./run_docker.sh