Skip to content

Latest commit

 

History

History
57 lines (42 loc) · 1.71 KB

01-Base-Images.md

File metadata and controls

57 lines (42 loc) · 1.71 KB

YOU DON'T NEED TO EXECUTE THESE STEPS IF LAB WAS CORRECTLY DEPLOYED

Notebook Images

Create a new project to put our analytics stuff:

oc new-project my-analytics

Create UBI openshift-spark-py36-inc build

oc new-build --name=openshift-spark-py36-inc https://github.com/mmgaggle/openshift-spark#wip-ubi \
             --context-dir=openshift-spark-build-inc-py36 \
             --strategy=docker

Watch openshift-spark-py36-inc build

oc logs -f bc/openshift-spark-py36-inc

Create openshift-spark-py36 build

oc new-build --name=openshift-spark-py36 \
             -i openshift-spark-py36-inc:latest \
             -e SPARK_URL=http://mmgaggle-bd.s3.amazonaws.com/spark-2.3.2-bin-hadoop-2.8.5.tgz \
             -e SPARK_MD5_URL=http://mmgaggle-bd.s3.amazonaws.com/spark-2.3.2-bin-hadoop-2.8.5.tgz.md5 \
             --binary
oc start-build openshift-spark-py36

Watch openshift-spark-py36 build (around 7 min)

oc logs -f bc/openshift-spark-py36

Create Jupyter Notebook build

Using the openshift-spark-py36 image stream as a base, we'll create a jupyter notebook build. The resulting image stream can be utilized by the jupyterhub operator when provisioning notebooks, and will ensure the notebooks have the correct spark client side library versioning necessary to interact with spark clusters provisioned by the spark operator with the openshift-spark-py36 image stream.

oc new-build --name=jupyter-notebook \
             https://github.com/luisrico/analytics-ml-lab \
             --context-dir=notebook \
             -i openshift-spark-py36:latest \
             --strategy=docker

Watch notebook build (around 18 min)

oc logs -f buildconfig/jupyter-notebook