hadoop connection question

[a212830]

Hi,

I'm testing out Hadoop Connect (and, hopefully, Shuttl). In Hadoop Connect, I get an error when attempting to add a hdfs cluster. It says "nvalid HADOOP_HOME. Cannot find Hadoop command under bin directory HADOOP_HOME='blahblahblah'. I checked with the Hadoop folks, and they have assured me that the Hadoop command exists in that structure. Am I missing something? Looking at the pop-up, it almost looks like the app is expecting Hadoop to exist on the Splunk server as well?

[bosburn_splunk · Accepted solution]

A212830,
You will need to install the hadoop CLI first - http://docs.splunk.com/Documentation/HadoopConnect/latest/DeployHadoopConnect/HadoopCLI. You can test to make sure it's working by signing in as the Splunk user and attempting to run the hadoop command.

[bosburn_splunk]

Everything is driven via the search head...

[a212830]

Thanks. Another question on Hadoop Connect - how does it work in a distributed environment? Do I need to put it on each indexer, or is everything driven from the search-head?