Peterlcc
commented
5 years ago Sorry for my negligence,i found i lost the core-site.xml in myconf directory.i fixed it immediately,but things are not going well.it still failed.When i look the stderr file.It only shows: ./start-component.sh: line 299: 2284 Terminated sleep 1
i'm lost again,so i try the other direction. i added the option "-namenode_servicerpc_addr centos-node1:9000" to use the physical NameNode.it worked! The applicationMaster started! i think maybe something confguration about the relationship between simulated NameNode and simulated DataNode was lost.Did i ? Meantime i'm confused where the relationship was set?
xkrogen
start-dynamometer-cluster.sh command: ./bin/start-dynamometer-cluster.sh -conf_path /root/dynamometer/dynamometer0.1.7/myconf -fs_image_dir hdfs:///dyno/fsimage -block_list_path hdfs:///dyno/blocks -hadoop_binary_path /root/dynamometer/dynamometer0.1.7/hadoop-2.8.3.tar.gz
the console error info: 19/07/18 11:23:37 INFO impl.YarnClientImpl: Submitted application application_1563419715675_0002 19/07/18 11:23:38 INFO dynamometer.Client: Track the application at: http://centos-node1:8088/proxy/application_1563419715675_0002/ 19/07/18 11:23:38 INFO dynamometer.Client: Kill the application using: yarn application -kill application_1563419715675_0002 19/07/18 11:23:58 INFO dynamometer.Client: NameNode can be reached via HDFS at: hdfs://centos-node2:9002/ 19/07/18 11:23:58 INFO dynamometer.Client: NameNode web UI available at: http://centos-node2:50077/ 19/07/18 11:23:58 INFO dynamometer.Client: NameNode can be tracked at: http://centos-node2:8042/node/containerlogs/container_1563419715675_0002_01_000002/root/ 19/07/18 11:23:58 INFO dynamometer.Client: Waiting for NameNode to finish starting up... 19/07/18 11:24:09 INFO dynamometer.Client: Infra app exited unexpectedly. YarnState=FINISHED. Exiting from client. 19/07/18 11:24:09 INFO dynamometer.Client: Attempting to clean up remaining running applications. 19/07/18 11:24:09 ERROR dynamometer.Client: Application failed to complete successfully
then i use "$?" to test the state of last commond : [root@centos-node1 dynamometer0.1.7]# echo $? 2
Inspired by the previous question,i check the application's stderr cat: metricsTailPIDFile: No such file or directory ./start-component.sh: line 299: 2207 Terminated sleep 1
and check the application's stdout: starting namenode, logging to /export/server/hadoop-2.8.3/logs/userlogs/application_1563419715675_0002/container_1563419715675_0002_01_000002/hadoop-root-namenode-centos-node2.out log4j:WARN No appenders could be found for logger (org.apache.hadoop.util.Shell). log4j:WARN Please initialize the log4j system properly. log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info. Going to attempt to upload metrics to: hdfs://centos-node1:9000/user/root/.dynamometer/application_1563419715675_0002/namenode_metrics Unable to upload metrics to HDFS Started namenode at pid 2124 Waiting for parent process (PID: 2005) OR namenode process to exit Cleaning up namenode at pid 2124 Deleting any remaining files
If i have something wrong in configuration?