unbridled-41 commented on PR #4618:
URL:
https://github.com/apache/rocketmq-dashboard/pull/4618#issuecomment-5760689258
@lizhimins Thanks for the review — the shared-flag concern was real. Fixed
in bc83c516 (branch rebased onto the current trunk `29a7d4ab`).
**Root cause**: the two registry loaders run concurrently (independent on
mount, fired together by the refresh button) but wrote one shared
`registryLoadError` boolean, so a success on either loader cleared the failure
the other had just recorded — e.g. nameserver success after a cluster failure
hid the banner while the Broker/Proxy tables kept stale/empty data.
**Review addressed**:
- Split the flag into per-loader error state: `registryClusterLoadError` /
`nameserverLoadError`. The banner renders while either is set; a successful
reload clears only its own error; "keep previous rows on failure" is unchanged.
- Both mixed-outcome orders are covered: cluster-fail + nameserver-success
and nameserver-fail + cluster-success, plus a partial-recovery case (one loader
recovers, the other still failing → banner stays; only when the failing side's
own reload succeeds does it clear).
- All 4 tests verified red against the trunk shared-flag source (`Unable to
find role="alert"` — the success cleared the other's error), green with the
fix: ClusterPage **33/33**; `tsc -b`, eslint, `npm run build` clean.
Details in the PR description ("Per-loader error state" section). Ready for
re-review.
--
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]
