allthingssecurity opened a new pull request, #29367: URL: https://github.com/apache/camel/pull/29367
# Description [CAMEL-26603](https://issues.apache.org/jira/browse/CAMEL-26603) With `forceCompletionOnStop`, the messages that a consumer still routed while its route stopped were lost. When a route stops, `DefaultShutdownStrategy` first prepares the route services, so the aggregator completes its groups, and then waits for the inflight and pending exchanges. Consumers that defer their shutdown (seda, direct and others) keep routing during that wait; seda drains its queue. Those messages started new groups. The strategy prepared the services again only after it had stopped the deferred consumers, but stopping a `DefaultConsumer` stops its processor, the route, so the aggregated exchange of these groups failed with `RouteStoppingException`, logged at WARN by the aggregator, and the messages were lost although they had been accepted (and taken off the seda queue). Found with Fray (controlled concurrency testing for the JVM): two threads sending to an aggregator with `forceCompletionOnStop` while a third stops the route; all sends succeeded, but only the first group came out. Change (`DefaultShutdownStrategy.ShutdownTask`): after the deferred consumers have been prepared for the shutdown (a seda consumer has then drained its queue and its threads have ended), and before they are stopped, the route services are prepared again (not forced), so the aggregator completes the groups of these messages while the route can still process them; it waits for its in-progress completions there. Not done for a suspend (the consumers are only suspended, the routes keep running) or after a timeout (the task returns, also when the timeout happened while the deferred consumers were prepared, such as a seda consumer draining its queue), and a forced shutdown passes `forced=true` as the neighbouring calls do. The other services that implement `prepareShutdown` (error handlers, loop, wire tap, on completion, JTA error handler, backlog debugger) only set a flag or do nothing, so a second call changes nothing for them. Not covered: a message that a deferred consumer such as direct still accepts after this point (between the second prepare and the stop of the consumer) is completed by the last prepare, as before, which fails. And when the whole CamelContext stops, an aggregated exchange sent to a seda queue whose route has already drained it stays in that queue. Tests: `AggregateForceCompletionOnStopLateMessagesTest` (one group aggregated, one message in flight, two in the seda queue, then `stopRoute` or `CamelContext.stop()`; latches, no sleeps). On main both fail (only `[m0]` comes out); with the fix both pass, and 20 runs each. camel-core aggregator package, shutdown strategy, graceful shutdown, abort after timeout, suspend/resume, route policy, route controller, seda and direct tests: 535 run, 0 failures apart from `SedaConfigureTest.testConfigureIgnoreCaseLocaleIndependent`, which fails because my local camel-package-maven-plugin is older than CAMEL-26439 and regenerated the configurers (not part of this change). # Target - [x] I checked that the commit is targeting the correct branch (Camel 4 uses the `main` branch) # Tracking - [x] If this is a large change, bug fix, or code improvement, I checked there is a [JIRA issue](https://issues.apache.org/jira/browse/CAMEL) filed for the change (usually before you start working on it): [CAMEL-26603](https://issues.apache.org/jira/browse/CAMEL-26603). # Apache Camel coding standards and style - [x] I checked that each commit in the pull request has a meaningful subject line and body. - [ ] I have run `mvn clean install -DskipTests` locally from root folder and I have committed all auto-generated changes. (I built the core modules and camel-core with the formatter and import-sort plugins and ran the related camel-core tests; no generated file changes. I did not run the full root build.) # AI-assisted contributions - [x] If this PR includes AI-generated code, commits have proper co-authorship attribution (e.g., `Co-authored-by` trailers) and the PR description identifies the AI tool used. This PR was prepared with Claude Code (Claude Opus 5.5). The commit carries a `Co-Authored-By` trailer. _Claude Code on behalf of allthingssecurity_ 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- 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]
