Update default node pool without re-installing k8s?

[dev-samples]

Based on this post:

https://pumpingco.de/blog/modify-aks-default-node-pool-in-terraform-without-redeploying-the-cluster/

It seems terraform does not support updating the default node pool (e.g. change VM Series, increase disk size etc.) without re-installing the whole cluster unless some az cli hacks are executed.

Is this still necessary or is there some documentation describing how to update the default node pool using terraform without using az cli?

Also based on this post:
#5311 (comment)

It seems that it not possible in AKS so what would the exact steps be to avoid downtime/loss of exsting workload/content?

I am currently experimenting with a variation/simpler version of:
https://pumpingco.de/blog/modify-aks-default-node-pool-in-terraform-without-redeploying-the-cluster/

1. Add new System node pool using az cli with new/required configuration (As I understand at least one active System pool needs to be active at all times to avoid a re-install of k8s that would remove all namespaces, workload etc.).
2. Delete old System node pool using az cli.
3. Crossing my fingers that all workload with automatically be migrated to the new System node pool.

I have also tried step 1 and 2 above using terraform but it results in a re-install of k8s (all workloads/namespaces are lost).

Regarding step 1 and 2 above I guess those changes somehow needs to be brought back into tfstate?

Any input is appreciated :-)

[tombuildsstuff]

hi @dev-samples

Thanks for opening this issue.

Support for cycling the Default Node Pool isn't currently available in the Provider and is being tracked in #7093 which unfortunately is blocked on the AKS API at this time, as we require the upstream issue Azure/AKS#1972 to be resolved to cycle the Default Node Pool - as such the Azure CLI or similar would work in the interim.

This forum is intended to be used for feature enhancements and bugs in the Azure Provider - so that we can keep this forum focused on that we instead ask that broader questions are raised using one of the Community Resources. As such I'm going to close this issue for the moment, but I believe you should be able to get an answer for this using one of the Community Resources.

Thanks!

[github-actions]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.