unbridled-41 opened a new issue, #4567:
URL: https://github.com/apache/rocketmq-dashboard/issues/4567
## Problem
`RocketMQLiteTopicProvider.isSupported()` probes only the first broker
master returned by `examineBrokerClusterInfo`:
```java
List<String> masters = masterAddresses(admin);
return !masters.isEmpty() && admin.getBrokerLiteInfo(masters.get(0)) != null;
```
If that single master runs a pre-lite build (mixed-version cluster / rolling
upgrade) or is momentarily unreachable, `getBrokerLiteInfo(masters.get(0))`
throws, the exception is swallowed, and the console reports `{"supported":
false}` — even when every other master in the cluster fully supports the lite
admin RPC. The LiteTopic page then degrades to "not supported" for the whole
cluster, and a transient failure of exactly one broker (the one that happens to
be enumerated first) flips the capability off with no retry.
The iteration order of `examineBrokerClusterInfo` is arbitrary from the
operator's point of view, so this is a single point of failure for the
capability signal.
## Evidence
-
`server/src/main/java/org/apache/rocketmq/studio/provider/apache/RocketMQLiteTopicProvider.java:110-113`
@ master `d50ffecc` — probe reads `masters.get(0)` only.
- Repro (unit level, two masters): `getBrokerLiteInfo(m1)` throws,
`getBrokerLiteInfo(m2)` returns a body → `isSupported()` currently returns
`false` although m2 answers.
## Expected behavior
Probe the masters and report supported when any of them answers the lite
admin RPC; a per-master probe failure just excludes that master. This matches
the same file's per-master degradation on every read path and the console's
cluster-wide semantics (the list page itself aggregates across masters).
## Related work
- #4565 / fix PR (open, same file): one failing master fails the list/quota
pages — the same single-master assumption on the read paths; this issue is its
capability-probe counterpart.
- PR #4562 (open, same file): per-master backlog aggregation in
`buildSummary`.
## PR
Fix incoming.
--
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]
