tju-yxq opened a new issue, #2597: URL: https://github.com/apache/rocketmq-dashboard/issues/2597
## Problem The Cluster page already coordinates instance-cluster polling carefully, but its auxiliary inventories use independent local `cancelled` flags: - NameServer registry entries - registry-discovered clusters used by the Proxy and Broker tabs - K8s certificate options Each effect has its own `cancelled` flag, so a slow earlier response can resolve after a newer refresh has started and overwrite the newer result. This is especially visible when the global Refresh button triggers `loadNsRegistry()` and `loadRegistryClusters()` while older requests are still in flight. Registry-cluster loading also sets loading/error/data state without request sequencing. ## Expected behavior - Sequence NameServer registry requests with a request generation. - Sequence registry-cluster requests with a separate request generation. - Sequence K8s certificate option requests. - Ignore stale success, failure, and loading updates. - Preserve the existing Refresh, create/update/delete, tab, and instance-cluster polling behavior. ## Verification scope Tests should cover stale registry-cluster responses being ignored after a refresh, while existing instance polling, refresh coordination, and registry mutation behavior remains unchanged. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
