canonical / canonical/data-science-stack
`create` should override existing service
- Dominant language
- Python
- Stars
- 37
- Forks
- 10
- PR merge metrics
- No merged PRs in 30d
Description
### Why it needs to get done
This change is affecting edge case:
When user is trying to create a notebook wile the service with the same name exists the dss will not allow that. Steps to reproduce.
```
pip install .
export DSS_KUBECONFIG=~/.kube/config # assuming config is set
dss initialize
kubectl create service clusterip empty-service --tcp=80:9376 -n dss
dss create empty-service --image=pytorch
```
output:
```
2024-04-19 11:55:53 [INFO] [main] [create_notebook_command]: Executing create command
2024-04-19 11:55:53 [ERROR] [create_notebook] [create_notebook]: Failed to create Notebook. Notebook with name 'empty-service' already exists.
Please specify a different name.
2024-04-19 11:55:53 [ERROR] [utils] [get_service_url]: Failed to get the URL of notebook name with error code 404.
Check the debug logs for more details.
2024-04-19 11:55:53 [DEBUG] [utils] [get_service_url]: Failed to get the URL of notebook name with error: services "name" not found
```
Instead of the existing behavior the dss create should override existing service.
### What needs to get done
`dss create` can create notebooks in dss namespace even if services with the same name already exist. These services will be overridden.
### When is the task considered done
Functionality is implemented and tested.
Contributor guide
Assessment
This issue has not been assessed yet.