Skip to content

Latest commit

 

History

History
65 lines (53 loc) · 3.71 KB

README.md

File metadata and controls

65 lines (53 loc) · 3.71 KB

Testing

The purpose of this document is to provide instructions for testing ADaPT-ML on your machine. Please make sure you have not modified the .env file. These tests will bring all the containers up using docker compose, test core functionality using the data for the example use case included in example_data, and will leave all the containers running so that you can view the test experiment runs through the MLflow UIs and/or investigate any issues using the containers' respective log files.

Unix: Run all tests

If you are using MacOS / Linux on your host machine, then run these commands:

cd ADaPT-ML/
bash ./test/all-unix.sh

Windows: Run all tests

  1. Open the Command Prompt in the ADaPT-ML root directory, and bring all the containers up:
docker-compose --env-file .env --profile dev up -d
  1. Make sure that none of the containers have exited or are restarting and wait until you see that the MLflow databases are "healthy", like this:
docker-compose ps
NAME                      COMMAND                  SERVICE             STATUS              PORTS
crate-db                  "/docker-entrypoint.…"   cratedb             running             0.0.0.0:4200->4200/tcp
dp-mlflow                 "/bin/bash"              dp                  running
dp-mlflow-db              "/entrypoint.sh mysq…"   dp_db               running (healthy)   33060-33061/tcp
dp-mlflow-server          "mlflow server --bac…"   dp_web              running             0.0.0.0:5000->5000/tcp
label-studio-dev          "/bin/bash"              ls                  running
label-studio-web          "./deploy/docker-ent…"   ls_web              running             0.0.0.0:8080->8080/tcp
modelling-mlflow          "/bin/bash"              m                   running
modelling-mlflow-db       "/entrypoint.sh mysq…"   m_db                running (healthy)   33060-33061/tcp
modelling-mlflow-deploy   "/start.sh"              m_deploy            running             0.0.0.0:8088->80/tcp
modelling-mlflow-server   "mlflow server --bac…"   m_web               running             0.0.0.0:5001->5000/tcp
  1. Run the test batch file:
test\all-windows.bat

Optional Clean-up

If you wish to bring the containers down after testing, then:

docker-compose down

Troubleshooting

  • If the crate-db container's status shows Restarting, please check its logs, docker logs crate-db. If you see something like this in the logs: 
    [1] bootstrap checks failed
[1]: max virtual memory areas vm.max_map_count [65530] is too low, increase to at least [262144] by adding `vm.max_map_count = 262144` to `/etc/sysctl.conf` or invoking `sysctl -w vm.max_map_count=262144`
    then check that you completed the appropriate bootstrap check steps.
  • If after the docker-compose ... up -d command, an error indicating that there is no space left occurs while pulling the images, then try these steps:
    1. Check the amount of reclaimable space for images: docker system df
    2. If reclaimable is greater than zero, clean up unused containers, networks, images, and the build cache: docker system prune -a
    3. Try the testing procedure again from the beginning
    4. If there still is not enough space, consider relocating the root directory (Linux, Windows)

For all other concerns with ADaPT-ML's tests, especially if any fail, please create an Issue.