Siyao Meng created HDDS-16425:
---------------------------------
Summary: DataNode startup can hang indefinitely after an
initialization failure
Key: HDDS-16425
URL: https://issues.apache.org/jira/browse/HDDS-16425
Project: Apache Ozone
Issue Type: Bug
Components: Ozone Datanode
Reporter: Siyao Meng
Assignee: Siyao Meng
A DataNode can remain running without completing startup or registering with
SCM when initialization of a required local service fails, for example because
a Raft log cannot be read due to corruption. Other startup callers may wait
indefinitely, and failures occurring after the heartbeat timeout may lose their
original exception, making diagnosis difficult.
Proposed changes:
* Ensure initialization has a clear success or failure outcome and release
waiting callers when it fails.
* Log the complete cause of local initialization failures and shut down the
DataNode.
* Preserve late task completions across heartbeat timeouts.
* Keep transient SCM connection failures retryable.
* Add regression coverage for concurrent startup, delayed failures, slow
successful initialization, and startup with an unavailable SCM.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
