Siyao Meng created HDDS-17209:
---------------------------------
Summary: Intermittent failure in
TestOMBootstrap#testBootstrapFollower
Key: HDDS-17209
URL: https://issues.apache.org/jira/browse/HDDS-17209
Project: Apache Ozone
Issue Type: Sub-task
Components: test
Reporter: Siyao Meng
Assignee: Siyao Meng
{code}
[ERROR] Tests run: 1, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 44.33 s
<<< FAILURE! -- in org.apache.hadoop.ozone.om.TestOMBootstrap
[ERROR] org.apache.hadoop.ozone.om.TestOMBootstrap.testBootstrapFollower()[2]
-- Time elapsed: 44.29 s <<< ERROR!
java.util.concurrent.TimeoutException:
Timed out waiting for condition. Thread diagnostics:
Timestamp: 2026-10-21 04:24:31,393
...
at
org.apache.hadoop.ozone.om.TestOMBootstrap.assertLogCapture(TestOMBootstrap.java:201)
at
org.apache.hadoop.ozone.om.TestOMBootstrap.testBootstrapFollower(TestOMBootstrap.java:168)
{code}
The follower did log the expected line ({{2026-10-21 04:24:01,589 ...
om.OzoneManager (OzoneManager.java:installCheckpoint(4892)) - Install
Checkpoint is finished with Term: 1 and Index: 64}}), 30 seconds before the
wait timed out, but the {{LogCapturer}} never received it.
h3. Root cause
{{Log4j1Capturer}} reuses the layout instance of the root {{stdout}} appender
for its own {{WriterAppender}}. Log4j 1 {{PatternLayout}} is not thread-safe:
{{format()}} resets and fills a single shared {{StringBuffer}}, relying on the
owning appender's lock. With two appenders sharing one layout, a thread logging
through the capturer and another thread logging through {{stdout}} format into
the same buffer concurrently, so the captured text can be replaced by, or mixed
with, another event. The same run shows this on the console as two events
garbled into one line ({{[omNode-3-SstFilteringService#0] 2026-10-21
04:23:46,625 [omNode-3-InstallSnapshotThread] INFOINFO ...}}). This affects
every test that uses {{LogCapturer.captureLogs(...)}} while other threads are
logging.
h3. Fix
Give each {{Log4j1Capturer}} its own {{PatternLayout}} with the same conversion
pattern instead of sharing the default appender's instance.
- https://github.com/apache/ozone/actions/runs/39536954941/job/119025384824
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
