Autoscaler producing error when using older Juju API with newer Juju controllers
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Kubernetes Autoscaler Charm |
Fix Released
|
High
|
Stone Preston |
Bug Description
The autoscaler pod that runs as part of the autoscaler charm was experiencing an error when trying to create the juju manager struct that handles the scale up and scale down process. The specific error generated was:
juju_cloud_
This was due to a change in the Juju API in newer versions of controllers (the error was confirmed on a juju 2.9.38 controller).
The autoscaler binary was updated to use a newer version of the Juju API in the following PR:
https:/
Additionally the PR updated some errors to be fatal (such as errors encountered when creating the manager), so that the autoscaler would stop running instead of continuing on whilst producing an error in the log.
This new binary was pushed into rocks to be consumed by the autoscaler charm as an OCI image.
The autoscaler charm was updated to use this new image in this PR:
https:/
description: | updated |
Changed in charm-kubernetes-autoscaler: | |
importance: | Undecided → High |
summary: |
- Autoscaler pod crashing when using older Juju API with newer Juju + Autoscaler producing error when using older Juju API with newer Juju controllers |
tags: | added: backport-needed |
Changed in charm-kubernetes-autoscaler: | |
status: | Fix Committed → Fix Released |
Backported with https:/ /github. com/charmed- kubernetes/ charm-kubernete s-autoscaler/ commit/ 773b2780ee1c128 c9d95b842db3d0f 93c686922d