JiaLiangC commented on PR #1162: URL: https://github.com/apache/bigtop/pull/1162#issuecomment-1676568567
@iwasakims Thank you for your suggestion. The YARN Timeline Service v2 relies on HBase and has three deployment modes: In the first mode, YARN starts an embedded HBase instance. In the second mode, YARN uploads YARN files to the resource manager and then launches an HBase instance as a YARN service. The third mode involves YARN utilizing an external HBase installation. The property yarn.timeline-service.hbase.coprocessor.jar.hdfs.location isn't restricted to HDFS paths only. It can work with any path that HDFS can resolve, such as file:/// or s3:/. Its purpose is to make it convenient for the HBase JVM to load the JAR file. yarn FlowRunTableRW.java This coprocessor JAR is primarily used in the code for creating the FlowRunTable in YARN.  hbase HTableDescriptor.java   In Ambari, a local path is used, following this pattern: yarn_timeline_jar_location = format("file://{stack_root}/{version}/usr/lib/hadoop-yarn/timelineservice/{coprocessor_jar_name}"). This approach is the simplest; it avoids the need to upload the JAR to HDFS. While uploading is still an option, hardcoding the version number in the code can complicate future maintenance. This is why I introduced this symbolic link. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
