[ https://issues.apache.org/jira/browse/HDFS-17382?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17822552#comment-17822552 ]
ASF GitHub Bot commented on HDFS-17382: --------------------------------------- steveloughran commented on PR #6584: URL: https://github.com/apache/hadoop/pull/6584#issuecomment-1973158445 @dineshchitlangia I want to strip dependendencies from hadoop common, rather than add more, as they contaminate all applications downstream and just add more CVEs. Just as we are adding a hadoop-common-zookeeper, maybe we can add a hadoop-common-server where we add server-side stuff only. we may still need to stitch these into hadoop/common/lib for release builds, but they won't be exported from the hadoop-common POM as transitive dependencies. Also, I like to consider 3.3.x as feature complete and we should focus our feature dev on trunk with backports to branch-3.4. non-asf forks can pick what they want... > Add Apache Log4j Extras Library to Hadoop 3.3 for Enhanced Log Rolling > Capabilities > ----------------------------------------------------------------------------------- > > Key: HDFS-17382 > URL: https://issues.apache.org/jira/browse/HDFS-17382 > Project: Hadoop HDFS > Issue Type: Improvement > Affects Versions: 3.4.0 > Reporter: woosuk.ro > Priority: Minor > Labels: pull-request-available > Attachments: image-2024-02-25-00-46-08-841.png > > > In the current Hadoop 3.4 version, the system relies on Log4j 1.x for logging > purposes. This dependency limits the logging functionality, especially when > it comes to rolling log files such as 'hdfs-audit.log'. Rolling log files is > crucial for managing log size and ensuring logs are rotated out over time to > prevent excessive disk space usage. However, the Log4j 1.x version integrated > within Hadoop lacks the necessary capabilities to efficiently handle log > rolling. > This library extends logging capabilities, including more flexible and > configurable log rolling features. By deploying this library, we can enable > advanced rolling strategies such as time-based rolling, size-based rolling, > and compression of rolled logs, which are not supported by the default Log4j > 1.x setup in Hadoop. > The integration of Apache Log4j Extras into Hadoop will significantly improve > log management by allowing for more sophisticated and configurable log > rotation policies. This enhancement is particularly important for maintaining > system performance and reliability, as well as for compliance with log > retention policies. > Although there are plans to upgrade to Log4j 2 in the forthcoming Hadoop 3.5 > version, which will inherently solve these issues by providing enhanced > logging features, there is an immediate need to enable advanced log rolling > capabilities in the current and previous versions of Hadoop. -- This message was sent by Atlassian Jira (v8.20.10#820010) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org