lizhimins commented on PR #6251:
URL: 
https://github.com/apache/rocketmq-dashboard/pull/6251#issuecomment-6120396603

   Thanks — this is merged, as part of 
[`0ffdc4a1a`](https://github.com/apache/rocketmq-dashboard/commit/0ffdc4a1afcdc7548ad4d607674422157494d6af)
 on `rocketmq-studio`.
   
   Your change went in together with #6173 — all yours, all in the same area — 
so we folded them into a single commit instead of landing several 
near-identical entries for the same subsystem. Nothing was dropped.
   
   The commit message calls this one out explicitly:
   
   > #6251 both `resendMessages` overloads run `validateResendTargetTopic` only 
when the caller passed a `targetTopic` (`RocketMQDLQProvider.java:202-203` for 
the time-range entry, `:274-275` for the selected-messages one), but when it is 
absent `resolveTargetTopic` (`:766-782`) infers the destination from 
`PROPERTY_DLQ_ORIGIN_TOPIC` or `PROPERTY_RETRY_TOPIC` and `resendOne` hands 
that name to a producer routing by topic across the whole endpoint — so a 
sibling cluster sharing the NameServer silently received the replayed batch 
while the API and the audit record reported success, which is exactly the 
scenario `requireSameClusterAsDlqTopic`'s own javadoc (`:714-725`) says must be 
rejected → `resendAll` (`:573-578`), the single point both entry points funnel 
through, now collects the inferred destinations into a `LinkedHashSet` and runs 
the existing `requireSameClusterAsDlqTopic(instanceId, %DLQ%<group>, 
destination)` on each one before `executeProducer`, rejecting the whole batch 
 with `BusinessException(409)`; a message whose destination cannot be resolved 
still fails individually as before, and a destination whose route cannot be 
resolved is left to the send rather than being reported as cross-cluster, 
matching the javadoc's rule that only a resolvable route can prove a mismatch.
   
   Closing as folded, not rejected.
   


-- 
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]

Reply via email to