ErrorStatusHandler - support for better error reporting in status #685
Conversation
| * common that the custom resource is changed during an execution, and it's much cleaner to have | ||
| * to original resource in place for status update. | ||
| */ | ||
| private R cloneResourceIfErrorStatusHandlerIfCouldBeCalled(R resource, Context context) { |
There was a problem hiding this comment.
This method name doesn't make sense :)
There was a problem hiding this comment.
:D
will rename it :)
There was a problem hiding this comment.
done. :) not sure if much better
|
|
||
| import io.fabric8.kubernetes.client.CustomResource; | ||
|
|
||
| public interface ErrorStatusHandler<T extends CustomResource<?, ?>> { |
There was a problem hiding this comment.
Once we have the HasStatus (or similar) interface, this probably should use that instead of CustomResource
There was a problem hiding this comment.
It will be moved to HasMetada if the other PR is merged.
…ava-operator-sdk into error-status-handler
| * common that the custom resource is changed during an execution, and it's much cleaner to have | ||
| * to original resource in place for status update. | ||
| */ | ||
| private R cloneResourceIfErrorStatusHandlerIfNeeded(R resource, Context context) { |
There was a problem hiding this comment.
The more I think about it, the less I'm sure we need to clone before updating the status. Presumably, if the resource has changed, the reconciliation will happen again and this update should be cancelled, anyway?
There was a problem hiding this comment.
as mentioned in the docs, it is for the case when the reconciliation executed but at some point an exception thrown, meannwhile in the logic of reconcile the status could be updated on the CR. I don't think we should use that, rather the original version of the resource. Since also info could be deleted, it's probably better to work with the fresh copy.
There was a problem hiding this comment.
Not convinced but I'll trust your judgement on this. :)
See linked issue.