metabase / druid-docker Goto Github PK

metabase/druid

docker build -t metabase/druid:0.20.2 .

The build logic ingests the data in rows.json by executing the ingestion spec task task.json. This is done in the script ingest.sh; tweak as needed.

Why ingest data as part of the build process? In some cases ingestion and indexing can take 10 minutes, on top of using an obnoxious amount of memory. Better to do it during build so we can use the image right out of the box instead of making CI super slow.

docker run -p 8081:8081 -p 8082:8082 -p 8888:8888 -it metabase/druid:0.20.2

For running Metabase tests you shouldn't need to change any of these.

• CLUSTER_SIZE -- Druid config to use. Currently one of nano-quickstart, micro-quickstart, small, medium, large, or xlarge. Default: nano-quickstart
• START_MIDDLE_MANAGER -- whether to start the middle manager process. Default false, because the middle manager is only needed for ingesting rows. Set to true if you plan to ingest more data.
• ENABLE_JAVASCRIPT -- whether to enable javascript on the Druid cluster. Metabase requires this, so by default it it true. Set it to something besides true to disable it.
• LOG4J_PROPERTIES_FILE -- Log4j2 config. By default, /druid/apache-druid-0.20.2/log4j2.properties, copied when building the Docker image, but you can mount a directory and supply a different file if you want different logging levels.

docker push metabase/druid:0.20.2