You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When implementing #1949 we considered how wether our new implementation is backwards compatible with 0.14.0. We have managed to migrate a 0.14.0 deployment to main by the following steps:
Push an app and bind it to some service
Stop korifi-controllers deployment
Delete the servicebindingio ServiceBindings for the app
Uprade Korifi
Restart the app (to recreate the app workload)
As deleting the servicebindingio binding should take place while korifi controllers are down, we could implement those steps in a pre-update helm hook
The manual app restart step is a bit tedious. In order to avoid it could we delete the AppWorkload while korifi controllers are down? Thus when the controllers are upgraded, they would reconcile the apps and create the app workload with desired bindings
The text was updated successfully, but these errors were encountered:
danail-branekov
changed the title
Improve a pre-upgrade helm hook to migrate existing service bindings
Implement existing service bindings migration
Feb 19, 2025
- It will delete the redundant sbio bindings, making sure they are
refclected in the app status AND restart the app so that it sees the
new environment
- The migration can be disabled
Co-authored-by: Yusmen Zabanov <[email protected]>
fixes#3820
- It will delete the redundant sbio bindings, making sure they are
refclected in the app status AND restart the app so that it sees the
new environment
- The migration can be disabled
Co-authored-by: Yusmen Zabanov <[email protected]>
fixes#3820
Co-authored-by: Yusmen Zabanov <[email protected]>
When implementing #1949 we considered how wether our new implementation is backwards compatible with 0.14.0. We have managed to migrate a 0.14.0 deployment to main by the following steps:
korifi-controllers
deploymentServiceBinding
s for the appAs deleting the servicebindingio binding should take place while korifi controllers are down, we could implement those steps in a pre-update helm hook
The manual app restart step is a bit tedious. In order to avoid it could we delete the
AppWorkload
while korifi controllers are down? Thus when the controllers are upgraded, they would reconcile the apps and create the app workload with desired bindingsThe text was updated successfully, but these errors were encountered: