6.4. Known Issues for HDFS

BUG-10264: Could not complete file error while running load generator on a HA cluster.

Problem: When a client sends a request to allocate a block to write to and the NameNode fails over at that time, client can fail to write to the file. The client is stuck permanently trying to recover from this error condition. This condition should rarely occur given failover must occur right when the addBlock operation is in progress.

Workaround: This bug has been fixed in Apache as part of HDFS-5257.