1

I am using CDH 5.4 with Pig 0.12. I am getting a lot of this error from all nodes:

TaskAttempt killed because it ran on unusable nodename:portnumber Container released on a *lost* node

What does this mean? In particular what does "lost" mean here? It doesn't look like the node is really lost in the cluster. Another question (more important question) is how to resolve this issue. Any help would be appreciated.

kee
  • 10,969
  • 24
  • 107
  • 168

1 Answers1

1

This particular case turned out to be a data storage problem. I restarted datanode manager from nodes which were lost with the message of "1/1 local-dirs are bad: /data/hadoop/yarn/local;"

kee
  • 10,969
  • 24
  • 107
  • 168