Open misohu opened 8 months ago
Thank you for reporting us your feedback!
The internal ticket has been created: https://warthogs.atlassian.net/browse/KF-5171.
This message was autogenerated
After discussion with the team we decided to go with design where we don't use juju and charms to deploy notebook-server and MLflow stack. In the newest design DSS components are deployed with plain Kubernetes objects (deployments + services + pvcs). DSS will also deploy MLflow with local mode. This means we no longer need MySQL and MinIO as everything will be stored in the folders.
The initialize command will also accept optional --default-image
parameter which may specify image to be used for the default user-notebook.
Because of this dss initialise
will:
--kubeconfig
parameter notebook-data
pvc which will be shared across notebooks created by user user-notebook
jupyter server with this image (if not --default-image
provided) which will accessible by user. This notebook will have notebook-data
pvc and it will have all environments variables needed to talk to MLflow user-notebook
and URL for accessing MLflow-ui.User can run
dss initialise --kubeconfig=/path/to/kubeconfig
with output:
Access the notebook at http://10.152.183.223/notebook/user-namespace/user-notebook/
Access MLflow ui at: http://10.152.183.34:5000
the above URLs are accessible by user and user can talk to MLflow server from inside the notebook.
User can override --default-image
parameter in dss initialise
to change the default notebook image used for DSS notebook.
Why it needs to get done
DSS needs a command to connect Juju (from within the snap) to Microk8s running on the local machine. For this task we may assume that both executables are available at local machine.
What needs to get done
Implement a CLI command
dss initialise
which connects juju to microk8s based on config speciffied in env variableKUBECONFIG
or based on flag--kubeconfig
. Initialisation will consist of:Example bash script from demo:
When is the task considered done
dss initialise
setups juju cloud controller and model and deploys DSS bundle.juju status
and we can see all the components of DSS bundle in active state.