We are running Hunk/Splunk 6.3.1 with Hive.
We saw some tasks for Hunk jobs failing due to no space left on device error. Our HDFS experts recommending setting
nfs.dump.dir = /tmp/.hdfs-nfs
I am testing this by setting
vix.nfs.dump.dir = /tmp/.hdfs-nfs
I wondered if others had run into this issue
As far as I can tell, what you are trying to do should work.
However, see this Hadoop bug regarding nfs.dump.dir: https://issues.apache.org/jira/browse/HDFS-8274
Therefore, you will need to make sure that the Hadoop version you are using already includes this patch.
As far as I can tell, what you are trying to do should work.
However, see this Hadoop bug regarding nfs.dump.dir: https://issues.apache.org/jira/browse/HDFS-8274
Therefore, you will need to make sure that the Hadoop version you are using already includes this patch.