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
Starting with the Operator SDK 0.12.0 (controller-runtime 0.3.0), the Jaeger Operator shows messages like the following during bootstrap:
WARN[0000] failed to upgrade managed instances error="the cache is not started, can not read objects"
ERRO[0000] error cleaning orphaned deployment error="the cache is not started, can not read objects"
This is caused by a change in the controller-runtime, which returns an error when the cache hasn't been started yet. Before this change, the client would just silently fail, meaning that our upgrade mechanism and the routine to clean orphaned deployments wasn't working properly.
To suppress the messages, we downgraded to 0.11.0 in #808, but a proper solution is to run the procedures above only when the cache has started. Determining when it happened is tricky, and it was suggested in the operator-sdk slack channel that we should make use of API clients instead.
The text was updated successfully, but these errors were encountered:
Starting with the Operator SDK 0.12.0 (controller-runtime 0.3.0), the Jaeger Operator shows messages like the following during bootstrap:
This is caused by a change in the controller-runtime, which returns an error when the cache hasn't been started yet. Before this change, the client would just silently fail, meaning that our upgrade mechanism and the routine to clean orphaned deployments wasn't working properly.
To suppress the messages, we downgraded to 0.11.0 in #808, but a proper solution is to run the procedures above only when the cache has started. Determining when it happened is tricky, and it was suggested in the operator-sdk slack channel that we should make use of API clients instead.
The text was updated successfully, but these errors were encountered: