2

We resize Amazon EMR cluster nodes from console.

When we added core nodes to the cluster, BlockMissingException occurs for a few /user/oozie/share/lib/ jars.

Replication factor for this /user/oozie/share/lib/ is 3 while default replication factor is 1.

Initially the cluster had 3 core nodes but when I increased it to 5 core nodes then some jars went missing. I couldn't understand why jars went missing when increasing nodes. Can anyone explain?

John Rotenstein
  • 241,921
  • 22
  • 380
  • 470
Pooja Soni
  • 137
  • 1
  • 2
  • 9
  • 1
    Adding a node to HDFS does not do anything to existing data unless you rebalance or change replication behavior. Only new writes will go the newly added nodes. If this happens again, you might need to enable AUDIT logging on HDFS to get more info on what happened to a particular block. – jc mannem Jan 22 '18 at 20:04
  • I see you have posted a related (probably preceding) question [here](https://stackoverflow.com/questions/48107616). Consider linking both of your related questions together by referencing them (*edit questions* / *add comments*). It helps those looking for an answer / trying to answer to get a better perspective. – y2k-shubham Jun 19 '18 at 08:49

0 Answers0