[PR] fix(proxy): keep stale probe failure from restoring old nodes [rocketmq-dashboard]

["btlqql (via GitHub)" <gi...@apache.org>]

btlqql opened a new pull request, #5835:
URL: https://github.com/apache/rocketmq-dashboard/pull/5835

   ## Why
   `applyProxyHome` re-checks the request id after the health-probe (topology) call resolves, but the failure path falls straight through to `setProxyNodes` / `persistProxyAddress` without that check. When a newer load supersedes an older one and the older probe then rejects, the superseded request writes its stale base node list back over the newer result.
   
   ## What changed
   - Re-check the request id on the topology-failure path in `applyProxyHome` and bail out when superseded, mirroring the success path.
   
   ## Validation
   - New test `ProxyStaleTopology.test.tsx`: first load parks on a pending probe, a refresh completes with different nodes, then the old probe rejects — the refreshed list must stay. Fails before the fix, passes after.
   - Existing `Proxy.test.tsx` suite (16 tests) still passes: `npx vitest run src/pages/studio/__tests__/Proxy`


-- 
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: commits-unsubscribe@rocketmq.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

Re: [PR] fix(proxy): keep stale probe failure from restoring old nodes [rocketmq-dashboard]

["lizhimins (via GitHub)" <gi...@apache.org>]

lizhimins closed pull request #5835: fix(proxy): keep stale probe failure from restoring old nodes
URL: https://github.com/apache/rocketmq-dashboard/pull/5835


-- 
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: commits-unsubscribe@rocketmq.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

Re: [PR] fix(proxy): keep stale probe failure from restoring old nodes [rocketmq-dashboard]

["lizhimins (via GitHub)" <gi...@apache.org>]

lizhimins commented on PR #5835:
URL: https://github.com/apache/rocketmq-dashboard/pull/5835#issuecomment-6120329512

   > Thanks for the careful write-up. The defect this PR describes was already fixed on > `rocketmq-studio` by `579918565` (#5032, which folded in #5256): the ownership check was moved > out of the `try` block to *after* the `catch`, so `Proxy.tsx:134-135` now runs > `if (requestId !== loadRequestId.current) return false;` before `setProxyNodes` / > `persistProxyAddress` on both the success and the probe-failure path. Since your branch has been > merged with the latest trunk, the guard you added inside `catch` is an exact duplicate of the very > next statement (same condition, same `return false`) and is a behavioural no-op — I could not > reproduce "fails before the fix, passes after" against `5e4c39b05`. > > The test itself is worth keeping, though: `ProxyStaleTopology.test.tsx` does go red if the guard > from #5032 is removed, and `Proxy.test.tsx` currently has no stale/superseded coverage. If you > re-file it as a test-only PR (`test(proxy): cover stale probe failure`, dropping
  the `Proxy.tsx` > hunk and rewording the body as a regression lock for #5032 rather than a new fix), I'm happy to > take it. Closing this one as already fixed.


-- 
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: commits-unsubscribe@rocketmq.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org