The teamcity platform helps to schedule the execution of Continuous Evaluation(CE) tasks.

There is also a web monitor displaying some analysis information.

how to start the teamcity

Start the official docker task, load from a backup project configuration.

Run the commands to start a docker

docker run -it --name teamcity-server-instance  \
    -v $(pwd)/teamcity_runtime/data:/data/teamcity_server/datadir \
    -v $(pwd)/teamcity_runtime/logs:/opt/teamcity/logs  \
    -p 80:8111 \
    -d \
    jetbrains/teamcity-server

how to start the agent

Start a paddle gpu docker image (the current working image on AWS is https://hub.docker.com/r/chunweiyan/dev9/)

Run the following commands to start a docker

nvidia-docker run -it  --name=gpu-agent3 -p 8080:8080 -v $(pwd)/teamcity_runtime/agent:/workspace chunweiyan/dev9  /bin/bash

install dependencies

apt install mongodb python3 python3-pip
pip3 install xonsh numpy pymongo

the mongodb service should be started before

install the agent follow the official guide.

ATTENTION: the agent/work should be bind to a path outside the docker.

how to run web server

dependencies

apt install python3 python3-pip mongodb openjdk-8-jdk
pip3 install flask flask-cache pymongo

the mongodb service should be started before

run

cd web
python3 main.py