KDB.AI Version Upgrade
This section details the steps to upgrade the KDB.AI Version on Azure Machine Learning Studio.
Pre-requisites
Before you can start the upgrade process of KDB.AI on Azure Machine Learning Studio, ensure you have the following:
- An active account for Azure Machine Learning Studio.
- The correct role-based access control permissions - you need a Key Vault Administrator or Owner role in the Resource group tab, so you can view, edit, and create secrets.
- Optionally, check your current version of KDB.AI.
How to check your user permissions.
- Go to the Resource group where your KDB.AI instance is deployed and select Access Control (IAM).
- In the Check access tab, click View my Access, to view the level of access assigned to you.
- To Add a new role, select Add from the top panel.
Steps to upgrade
- Go to the Resource group where KDB.AI is deployed and access the Key vault.
- On the left panel, go to Secrets. Make sure you have a role that allows you to view, edit, and create secrets.
- Select crImageVersion from the list.
- Click on New Version from the top panel.
-
To upgrade your current version, you have to complete the following fields:
Field name Field value Upload options: Manual Secret value: input the version you want to upgrade to (for example 1.0.1) Content type (optional): text/plain Enabled: YES (if you would like to automatically update the running version) When completed, click on Create.
-
Now you need to restart the KDB.AI instance. To do so, go back to the Resource group and access the KDB.AI service.
- Click on the Launch studio button.
- From the left panel, access Compute.
- Tick your KDB.AI instance and press Restart.
- To verify if the upgrade is successful, follow the instructions below:
How to check your current version of KDB.AI.
How to check version
- Go to the Resource group where KDB.AI is deployed and access the ml-kdbai service.
- Click on the Launch studio button.
- Access Notebooks from the left panel.
- Go to the Logs folder and click on the stdout file to open it.
- Search for kdbai-db: (the set of digits next to it shows the KDB.AI version that's currently deployed).