Hi all,
I’ve been playing around with Metabase for a couple of weeks now and I really like where it is going.
I’m using docker with docker-compose to run my metabase instance, building it from source. The first time I built it, I set metabase to connect to a Postgres database (instead of the default H2) and added a data source to run questions against it - configured it’s metadata, added some cards and dashboards. The Postgres instance serving metabase config is also running on docker, persisting data to an host volume.
Starting and stoping the metabase instance seemed to be working like a charm as it would reconnect to it’s Postgres database containing all the questions and dashboards and resume in the same point it was before the shutdown.
Today, I updated to the latest changes on master and rebuilt the container for metabase and when I start metabase again, it’s pointing me to /setup/. When I connect to the Postgres database, all metabase tables are there with the data I previously created while running the pre-updated metabase instance which can only mean (I assume) metabase ignored it and must be using it’s default H2 database again.
My docker-compose.yam file hasn’t change and looks like this:
`metabase:
build: .
ports:
- 3000:3000
links:
- postgres
volumes:
- /tmp:/tmp
environment:
- MB_DB_TYPE=postgres
- MB_DB_DBNAME=metabase
- MB_DB_PORT=5432
- MB_DB_USER=postgres
- MB_DB_PASS=""
- MB_DB_HOST=postgres
postgres:
image: clkao/postgres-plv8:9.5
ports:
- 5432:5432
volumes:
- /var/lib/postgresql/data:/var/lib/postgresql/data
environment:
- POSTGRES_DB=metabase
`
I’m a bit lost because I was assuming every time a metabase instance is created, it should read all properties from the defined MB_DB_DBNAME.
Am I missing something? Does metabase store any important information on files in the defined docker volumes: (- /tmp:/tmp in my case) and it’s being wiped out when I restart my computer?
Can anyone help me out here?