0

All need some help regarding process getting killed in data node. Have a cluster with 1 name node and 5 data nodes. Have installed hive to run queries. Every time I run hive query I loose connection between the data node and the name node. This happens almost always. Is there any specific reason why this happens. Nothings gets written in the log file also. Any help in resolving this would be of great help. Please let me know where I should look for the issues as I am new to this area.

Thanks

Sunil
  • 131
  • 1
  • 9
  • Possible duplicate: https://stackoverflow.com/questions/21999192/hdfs-data-node-disconnected-from-namenode – Keshav Pradeep Ramanath Jun 09 '17 at 03:29
  • I am not using Cloudera. Also the connection is there. The mapper runs and then suddenly the data node dies out. It is not a port issue I can tell that for sure because randomly data nodes go down, – Sunil Jun 09 '17 at 03:33
  • Can you mention what is getting logged while the connection is lost? – Keshav Pradeep Ramanath Jun 09 '17 at 03:34
  • I have actually restarted the cluster and have run 7 to 8 jobs and all ran successfully without an data node loss. Will post the logs as soon as I encounter a data node loss. So please bear with me for the time till I encounter data node loss. By the way thanks for helping me out with this. – Sunil Jun 09 '17 at 03:45
  • Again started loosing the data nodes but unable to paste the log as I am exceeding the limit. Is there a way I can add the log ? – Sunil Jun 13 '17 at 08:06

0 Answers0