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 performing a sync on a KafkaCluster resource, if there are any broker configuration changes, the Kafka operator will perform a rolling update on all brokers and mark certain brokers with a state of "ConfigOutOfSync". During this process, ArgoCD will immediately report the resource in "Degraded" state until all brokers are updated, after which it goes into "Healthy" state.
I believe "Progressing" state would be more appropriate in this circumstance instead of "Degraded", since it is in the process of updating and not an error state.
In addition, this causes issues when performing a sync/wait operation from a workflow task since it will immediately return when it is in "Degraded" state and will not wait until the sync is healthy.
Below are the states of each broker and the KafkaCluster status during the rolling update:
Modify helm chart or values.yaml file to update the broker config for the KafkaCluster (such as adding/removing allow.everyone.if.no.acl.found: true from spec.readOnlyConfig)
Perform a sync.
Expected behavior
After invoking a sync, the KafkaCluster is in "Progressing" state while it is performing the rolling update to the brokers. After the rolling update is complete, the state changes to "Healthy".
Checklist:
Describe the bug
When performing a sync on a KafkaCluster resource, if there are any broker configuration changes, the Kafka operator will perform a rolling update on all brokers and mark certain brokers with a state of "ConfigOutOfSync". During this process, ArgoCD will immediately report the resource in "Degraded" state until all brokers are updated, after which it goes into "Healthy" state.
I believe "Progressing" state would be more appropriate in this circumstance instead of "Degraded", since it is in the process of updating and not an error state.
In addition, this causes issues when performing a sync/wait operation from a workflow task since it will immediately return when it is in "Degraded" state and will not wait until the sync is healthy.
Below are the states of each broker and the KafkaCluster status during the rolling update:
To Reproduce
allow.everyone.if.no.acl.found: true
fromspec.readOnlyConfig
)Expected behavior
After invoking a sync, the KafkaCluster is in "Progressing" state while it is performing the rolling update to the brokers. After the rolling update is complete, the state changes to "Healthy".
Screenshots
N/A
Version
Logs
N/A
The text was updated successfully, but these errors were encountered: