Skip to main content

Databases

The DSRI provides templates for several SQL, NoSQL, and graph databases. You can deploy them directly from the DSRI Catalog or via Helm.

Once a database is running, you can connect to it from another pod in the same project using the database service name as hostname. To list services in your project:

oc get services
info

Our cluster enforces mandatory CPU, memory, and ephemeral-storage requests/limits on every pod. All templates below already account for this, if you create a custom pod yourself (e.g. a test/client pod), remember to set these explicitly or it will be rejected.

SQL databases​

PostgreSQL​

Find and instantiate the PostgreSQL template in the DSRI Catalog. Username and password are auto-generated if left blank when instantiating.

To connect from another pod, first install the client:

apt-get update && apt-get install postgresql-client -y

Retrieve the generated credentials:

oc get secret postgresql -o jsonpath='{.data.database-user}' | base64 -d; echo
oc get secret postgresql -o jsonpath='{.data.database-password}' | base64 -d; echo
oc get secret postgresql -o jsonpath='{.data.database-name}' | base64 -d; echo

Then connect using the service name:

psql -h postgresql -U <username> -d <database-name>

See the dsri-demo repository for an example of accessing PostgreSQL from a Jupyter notebook.

MySQL​

Find and instantiate the MySQL template in the DSRI Catalog.

To connect from another pod, first install the client:

apt-get update && apt-get install mariadb-client -y

Retrieve the generated credentials:

oc get secret mysql -o jsonpath='{.data.database-user}' | base64 -d; echo
oc get secret mysql -o jsonpath='{.data.database-password}' | base64 -d; echo
oc get secret mysql -o jsonpath='{.data.database-name}' | base64 -d; echo

Then connect using the service name:

mysql -h mysql -u <username> -p<password> <database-name>

See the dsri-demo repository for an example of accessing MySQL from a Jupyter notebook.

tip

MySQL can also be deployed using Helm. See the Helm documentation for details.

NoSQL databases​

MongoDB​

MongoDB is a document-based distributed database. It can be deployed via Helm using the Bitnami chart.

First, install the Helm client as described in the Helm documentation. Then add the Bitnami repository:

helm repo add bitnami https://charts.bitnami.com/bitnami
helm repo update

Deploy MongoDB:

helm install my-mongodb bitnami/mongodb

Retrieve the root password:

export MONGODB_ROOT_PASSWORD=$(oc get secret --namespace my-project my-mongodb -o jsonpath="{.data.mongodb-root-password}" | base64 -d)

To connect from within the cluster, create a temporary MongoDB client pod (replace my-project with your project name):

oc run --namespace my-project my-mongodb-client --rm --tty -i --restart='Never' \
  --env="MONGODB_ROOT_PASSWORD=$MONGODB_ROOT_PASSWORD" \
  --image registry-1.docker.io/bitnami/mongodb:latest --command -- bash

Then connect to the database:

mongosh admin --host "my-mongodb" --authenticationDatabase admin \
  -u root -p $MONGODB_ROOT_PASSWORD

To connect from outside the cluster:

oc port-forward --namespace my-project svc/my-mongodb 27017:27017 &
mongosh --host 127.0.0.1 --authenticationDatabase admin \
  -u root -p $MONGODB_ROOT_PASSWORD

Redis​

Redis is a key-value cache and store, often used as a data structure server.

Find and instantiate the Redis template in the DSRI Catalog. A connection password is auto-generated for you.

Retrieve the password and connect from another pod in the same project using the service name as hostname:

export REDIS_PASSWORD=$(oc get secret redis -o jsonpath="{.data.database-password}" | base64 -d)

redis-cli -h redis -a $REDIS_PASSWORD ping

A successful connection returns PONG.

Graph databases​

Find and instantiate the Virtuoso triplestore template in the DSRI Catalog. The deployment is based on the open source version of Virtuoso, and runs on port 8890.

Connect from another pod in the same project using the service name as hostname:

curl -I http://virtuoso:8890/sparql

A successful connection returns HTTP/1.1 200 OK.

Ontotext GraphDB​

Find and instantiate the Ontotext GraphDB template in the DSRI Catalog. No manual build required. The template uses a pre-built image ready to deploy directly.

info

If you have an enterprise license, you can point the template's image parameter at the official DockerHub image instead. For the free edition, note that as of GraphDB 11, Ontotext requires requesting a separate free license before the database will start: see the GraphDB Free download page for details. The image used by default in this template predates that requirement.

Connect from another pod in the same project using the service name as hostname:

curl -I http://graphdb:7200/