Konstantin Bereznyakov created HDFS-17907:
---------------------------------------------
Summary: DataXceiver logs ERROR for "Premature EOF" on WRITE_BLOCK
Key: HDFS-17907
URL: https://issues.apache.org/jira/browse/HDFS-17907
Project: Hadoop HDFS
Issue Type: Bug
Reporter: Konstantin Bereznyakov
DataXceiver logs ERROR for "Premature EOF" exceptions during WRITE_BLOCK
operations. These are expected when clients disconnect during writes and should
be logged at INFO/DEBUG level instead.
[Affected
Code|https://github.com/apache/hadoop/blob/trunk/hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataXceiver.java#L309]
*Current Behavior*
ReplicaAlreadyExistsException during WRITE_BLOCK is already handled specially:
if (op == Op.WRITE_BLOCK && t instanceof ReplicaAlreadyExistsException)
{ // logs at INFO/TRACE, not ERROR }
But "Premature EOF" exceptions during WRITE_BLOCK fall through to ERROR
logging, even though they are equally expected when clients abort writes.
*Impact*
- Unnecessary ERROR-level log noise in DataNode logs
- Difficult to identify real errors among expected client disconnections
- Inconsistent with existing ReplicaAlreadyExistsException handling
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
