Hi everyone, I Have a general question, about how the mappers are assigned to a job that are running. I have looked for this in aws documents and apache hadoop sites for knowledge but still am little confused by seeing the error I am getting.
Scenario: So I was running Sqoop commands(75873280 jobs) in aws EC2 instance and I have 3data nodes and one master node. Each instance have 120GB storage capacity. And my data is ~10-20GB in total. Configurations with 1 replication, 8 mappers set in "mapreduce.tasktracker.map.tasks.maximum" and "mapreduce.job.maps" column in mapred-site.xml. and my sqoop command have number of mappers set to 4. When I start running these Sqoop job using script file it runs for ~60000 jobs and then I have been getting the below error: `Error: org.apache.hadoop.ipc.RemoteException(java.io.IOException): File /user/hive/warehouse/test.db/test_12km_p/_SCRATCH0.03130391930415266/time_stamp=2016-05-11/_temporary/1/_temporary/attempt_1525190353973_2740_m_000002_0/part-m-00002 could only be replicated to 0 nodes instead of minReplication (=1). There are 3 datanode(s) running and no node(s) are excluded in this operation.` And after some time, if I rerun the job, instead of restarting the hadoop deamons, it is executing perfectly for some time and throw the same error at somepoint again. Is there any permanent fix for this issue? Thanks for any suggestions.
