Starting Jupyter Notebook Server at 96%, Invalid Image Name
Environment
- OpenShift Cluster Platform 4.x
- Self Managed Red Hat OpenShift AI
- 1.2x
Issue
The user is not able to finalize the spawning of a Jupyter Notebook, the Server is stuck at 96% and the notebook pod returns a Warning: failed to apply default image tag ":py3.8-v1": couldn't parse image reference ":py3.8-v1": invalid reference format
Resolution
Red Hat OpenShift Data Science(RHODS) use pre-built notebooks and they will be pulled to the internal registry.
For this reason, you must make sure that the internal registry is running in the cluster before you install RHODS.
Verify if the cluster has an image registry pod
oc get pods -n openshift-image-registry
NAME READY STATUS RESTARTS AGE
cluster-image-registry-operator-759dbc577c-28qjt 1/1 Running 0 1d
image-registry-584cc85cf4-28nw6 1/1 Running 0 1d (<=== This is the image-registry)
node-ca-q4f98 1/1 Running 0 1d
If you don't see any image registry pod is not running, you must to deploy it. In order to deploy an image registry pod, please refer to the Openshift Container Platform Documentation.
For a test cluster, you can use emptyDir which is NOT recommended for a production cluster.
oc patch configs.imageregistry.operator.openshift.io cluster --type merge --patch '{"spec":{"storage":{"emptyDir":{}}}}'
config.imageregistry.operator.openshift.io/cluster patched
oc patch configs.imageregistry.operator.openshift.io cluster --type merge --patch '{"spec":{"managementState":"Managed"}}'
config.imageregistry.operator.openshift.io/cluster patched
Root Cause
At least, one image registry pod must run in an openshift container platform cluster. Without this, the cluster does not mean that it is ready to provide services properly.
Diagnostic Steps
- Install RHOAI self managed
- Log in as RHOAI user
- Spawn a notebook from the RHOAI Dashboard
- Check the notebook pod
$ oc get pods -n rhods-notebooks
Example output
NAME READY STATUS RESTARTS AGE
jupyter-nb-admin-0 1/2 InvalidImageName 0 4m48s
This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.
Comments