[
https://issues.apache.org/jira/browse/CAMEL-27091?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Work on CAMEL-27091 started by shashank.
----------------------------------------
> camel-mongodb-gridfs - each stop of the consumer leaves a thread waiting for
> ever for the polls
> -----------------------------------------------------------------------------------------------
>
> Key: CAMEL-27091
> URL: https://issues.apache.org/jira/browse/CAMEL-27091
> Project: Camel
> Issue Type: Bug
> Components: camel-mongodb-gridfs
> Reporter: shashank
> Assignee: shashank
> Priority: Major
>
> h3. What happens
> {{GridFsConsumer}} runs {{run()}} on a thread of its own executor. {{run()}}
> builds a {{BackgroundTask}} with an unlimited duration and calls
> {{task.run(...)}}, which schedules the polls on {{taskExecutor}} and waits
> until the task ends. A poll never ends the task (it returns false whatever it
> found, also when it fails).
> {{doStop}} shuts {{taskExecutor}} down, which cancels the scheduled polls,
> and shuts down the executor without interrupting it. Nothing releases the
> thread in {{run()}}: every stop of the consumer (route stop and start, route
> reload) leaves one thread waiting for ever, and the task stays in the task
> registry.
> h3. How it was found
> Found with a TLA+ model of the thread in {{run()}}, the poll thread and the
> stop, checked with TLC. On main the invariant "after the stop, no thread
> waits on a latch that nothing can count down" is violated in 25 states
> (start, polls, stop), and no behaviour of main reaches a stopped consumer
> whose {{run()}} has returned. Keeping {{task.run}} and cancelling the task in
> doStop fixes that but interrupts a poll that is routing files (the finally of
> {{waitForTaskCompletion}} cancels the future with interrupt). The fix below
> holds all three properties (no stranded thread, no interrupted poll, no task
> left registered; 52 distinct states).
> h3. How to reproduce
> {{GridFsConsumerTimestampTest.stopLeavesNoThreadWaitingForThePolls}} (fake
> bucket): start the consumer, wait for two polls, stop it: its thread is still
> in {{run()}} 20 seconds later.
> h3. Fix
> {{run()}} schedules the polls with {{task.schedule(...)}} instead of waiting
> for the task, keeps the task, and does not schedule when the consumer is
> already stopping; {{doStop}} cancels the task without interrupt, so a poll
> that runs finishes its files and the task leaves the registry.
> _Filed with Claude Code on behalf of allthingssecurity._
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)