iBenny created HIVE-25794:
-----------------------------

             Summary: CombineHiveRecordReader: log statements in a loop leads 
to memory pressure
                 Key: HIVE-25794
                 URL: https://issues.apache.org/jira/browse/HIVE-25794
             Project: Hive
          Issue Type: Bug
          Components: Logging
            Reporter: iBenny


Similar to HIVE-16150, a huge string will be built in a loop, even the log 
level is INFO. That leads to memory pressure when processing a big number of 
split files. 

>From 
>[CombineHiveRecordReader.java|https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/io/CombineHiveRecordReader.java#L116],
> the following needs to be fixed.



LOG.debug("Found spec for " + path + " " + otherPart + " from " + 
pathToPartInfo);



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

Reply via email to