An extendable Docker image for Airbnb's Superset platform.
- Python 3.6 + Debian Stretch
- AirBnB Superset
- PostgreSQL driver
- MySQL driver
- MS SQL/Sybase through ODBC
- Docker Compose setup
By default the Superset meta database will be stored in a local sqlite database, in the most basic case getting a working Superset instance up and running is as simple as:
docker run -d --name superset -p 8088:8088 tylerfowler/superset
The entrypoint script will set up an admin user for you using the ADMIN_*
environment variables, with a default username and password of:
username: admin
password: superset
MySQL is in separate container, Redis cache is enabled. Visit http://localhost:8088. If you have jwilder/nginx-proxy, http://superset.local will be available. Modify docker-compose.yml to your liking.
docker-compose -p superset up -d
The admin user is created in the entrypoint script using the ADMIN_*
environment variables in the Dockerfile, which should be overriden.
docker run -d --name superset \
-e ADMIN_USERNAME=myadminuser \
-e ADMIN_FIRST_NAME=Some \
-e ADMIN_LAST_NAME=Name \
-e ADMIN_EMAIL=nobody@nowhere.com \
-e ADMIN_PWD=mypassword \
-p 8088:8088 \
huksley/superset
In order to keep the base image as lean as possible only the Postgres driver is included and any other database drivers or libraries that are needed should be installed in a downstream image. To use a different backend you just need to install the appropriate drivers and modify the $SUP_META_DB_URI
to be the database connection string for the backend, which is only used in the entrypoint script at runtime.
The Superset config file is generated dynamically in the entrypoint script using the SUP_*
environment variables, for example to increase the row limit to 10000 and the number of webserver threads to 16:
docker run -d --name superset \
-e SUP_ROW_LIMIT=10000 \
-e SUP_WEBSERVER_THREADS=16 \
-p 8088:8088 \
huksley/superset
In order to correctly set up Superset the entrypoint needs to be set the superset-init.sh
script, though if a more advanced configuration is required you can also supply your own entrypoint script.
In your Dockerfile add any script as long as it ends up at /docker-entrypoint.sh
. This script will be run after the initial superset_config.py
is generated but before any of the Superset setup commands are ran. Note that the environment variables will still be used to bootstrap the Superset configuration file.
For example to add a Redis cache to your configuration:
#!/bin/bash
cat <<EOF >> $SUPERSET_HOME/superset_config.py
CACHE_CONFIG = {
'CACHE_TYPE': 'RedisCache',
'CACHE_REDIS_URL': 'localhost:6379'
}
EOF
After this is finished running Superset will continue to configure itself as normal. Alternately, if the init script detects that a superset-config.py
file already exists under $SUPERSET_HOME
then it will skip bootstrapping the file altogether and will use the user supplied config instead. Similarly after Superset is finished setting itself up (migrating the DB, initializing, creating admin user, etc...) it will write an empty file at $SUPERSET_HOME/.setup-complete
so that subsequent runs on a mounted volume will not set up Superset from scratch. To take advantage of this fact simply mount the $SUPERSET_HOME
directory (which is /superset
by default).
docker run -d --name superset \
-v /mysuperset:/superset \
-p 8088:8088 \
huksley/superset
Note, however, that even if an existing Superset configuration is detected, any user supplied docker-entrypoint.sh
file will still be run. So if need be write a file that can be checked for to ensure your script only runs once in the same fashion that the superset-init.sh
script does.
Enjoy!
After starting up, use these templates to specify URIs to databases
- mysql+mysqlconnector://user:password@host[:port]/dbname
- sybase+pyodbc://username:password@dsnname[/database]
- postgresql+psycopg2://user:password@host:port/dbname[?key=value&key=value...]
- Consult original docker image at https://github.com/tylerFowler/docker-superset for additional configuration.
- http://www.py-my.ru/post/50fe762371144856ca000006
- http://docs.sqlalchemy.org/en/rel_1_1/dialects/index.html