Dev Environment for Caravel on Drill
The goal is create an easy environment with all dependencies to work on getting caravel working with Apache Drill
Steps to Use Image
- Edit the builddocker.sh and run.sh files and edit the DOCKERIMG and PROJECTDIR variables to match your project (DOCKERIMG should match in both files)
- Run builddocker.sh - This builds the docker container
- Run run.sh This drops you into /app/caravel in the container. If this is the first time you've run things, run initial.sh to config things.
- Caravel will be running now! If you want to reuse config, run
caravel runserver -d
Some Caravel Notes
- We use host networking to make this easier. If you have something running on port 8088 on your docker host, it will conflict.
- After the initial config (initial.sh) the files for the caravel sqllite DB are created on the host. So even if you stop the image, your sqllite will be maintained.
- The initial.sh and caravel_config.py is copied into ./working (and thus /app/caravel) everytime you run run.sh. Thus if you want to start form scratch. Just stop the container and rm ./working/*
- If you make changes to caravel_config.py it will copy to to working each time you run it.
- I want to explore the Flask caching more, just haven't yet.
PyODBC Notes
- This container comes with the MapR ODBC Driver and PyODBC (with UnixODBC) Installed
- To use rename pyodbc_example.py.template to pyodbc_example.py
- Once renamed, updated the python to use your hostnames, select ZK or DIRECT, and then put in your credentials (for testing)
- If the file pyodbc_example.py exists, the run.sh will copy it to /app/caravel in the container for testing.
- This is just a basic demonstration that pyodbc, in the container is working.
SQL Alchemy Drill Dialect Notes
- We do not install the SQL Alchemy dialect in the Docker container (Because it't not complete or working)
- To run it, copy into the container (I use the mounted application location) and run "python setup.py install"
- This is NOT WORKING (hence why it is not in the container) It's a copy of the Access dialect, that is slowly morphing into a drill dialect.
- It's initially here to allow a connect. Basic things are being updated slowly as time goes on, it will allow a "test connection" but that's about it.