Share via

Migrate node image version to Azure Linux 3.0

Nagashree 80 Reputation points
2025-09-22T11:18:44.7733333+00:00

What is the rollback process if we need to revert after migrating to version 3.0?

If we run the following command to update the node pool:

az aks nodepool update \ --resource-group $RESOURCE_GROUP \ --cluster-name $CLUSTER_NAME \ --os-type Linux \ --os-sku AzureLinux3 \ --kubernetes-version 1.30.0 \ --name $NODE_POOL_NAME \ --node-count 1

Can we simply change the --os-sku parameter to AzureLinux2 and execute the command again to roll back?

I’ve checked the Microsoft documentation on upgrading OS versions but couldn’t find a direct answer.

How long will the upgrade process take for one node?

Azure Kubernetes Service
Azure Kubernetes Service

An Azure service that provides serverless Kubernetes, an integrated continuous integration and continuous delivery experience, and enterprise-grade security and governance.


Answer accepted by question author

Ankit Yadav 14,450 Reputation points Microsoft External Staff Moderator
2025-09-22T22:50:04.0433333+00:00

Please find answers to your queries asked earlier:

What is the rollback process if we need to revert after migrating to version 3.0? Can we simply change the --os-sku parameter to AzureLinux2 and execute the command again to roll back?

Yes you are almost right for the rollback process. You can use same command, but you'd be required to change --os-sku parameter to AzureLinux and that will pick the default OS version for the Kubernetes version your cluster is running onto.User's image

see references: https://learn.microsoft.com/en-us/azure/aks/upgrade-os-version#roll-back-your-os-version-to-the-default-os-sku

How long will the upgrade process take for one node?

There isn’t a set timeframe for when a node will be upgraded to the new OS version, it depends on each cluster, so the timing can vary.

Roughly, an OS upgrade on the node will undergo these steps

drain -> reimage -> uncordon

And this also varies on various factors such as how much drain timeout, surge values, heavy/lightweight workloads, stateful/large data are there on the cluster. Thus, we won't be able to comment on the duration for the upgrade process.

Let me know if this clarifies your queries!!

Was this answer helpful?

1 person found this answer helpful.
0 comments No comments

0 additional answers

Sort by: Most helpful

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.