Claus Ibsen created CAMEL-24195:
-----------------------------------
Summary: camel-aws2-s3 - StreamUploadProducer race condition and
timeout task exception handling
Key: CAMEL-24195
URL: https://issues.apache.org/jira/browse/CAMEL-24195
Project: Camel
Issue Type: Bug
Components: camel-aws
Reporter: Claus Ibsen
Split from CAMEL-24157 (medium-severity findings, group 3).
*StreamUploadProducer race condition and timeout task cancellation*
{{AWS2S3StreamUploadProducer:181-183}} — {{uploadAggregate}} is read and
mutated outside the lock in {{processWithoutTimestampGrouping()}}. The timeout
task ({{StreamingUploadTimeoutTask.run()}}) acquires the lock and can set
{{uploadAggregate = null}}. Race: main thread reads {{uploadAggregate !=
null}}, timeout task nulls it, main thread dereferences → NPE.
Additionally, {{StreamingUploadTimeoutTask.run()}} does not catch exceptions
from {{uploadPart()}}/{{completeUpload()}}. Since
{{ScheduledExecutorService.scheduleAtFixedRate}} silently cancels all future
executions when the task throws an uncaught exception, any S3 API error in the
timeout task permanently disables the timeout flush mechanism for the lifetime
of the producer.
Recent CAMEL-24153 fix addressed different bugs (maxRead clamping, orphaned
multipart) but did not address the lock-ordering race or exception handling.
*Fix approach:*
- Move the {{uploadAggregate}} reads at lines 181-183 inside the lock
- Wrap the timeout task body in try-catch to log and continue on S3 API errors
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
