Hadoop Client
Client machines have Hadoop installed with all the cluster
settings, but are neither a Namenode or a Datanode. The difference is that
client machine don’t have any hadoop services started. That’s why we are
able to run commands to interact with hdfs if we mount using fuse.
Besides this Clients can have components like Hive / Sqoop running from where users can interact with the cluster
For more details please see
https://www.opensciencegrid.org/bin/view/Storage/HadoopInstallation
No comments:
Post a Comment
Please share your views and comments below.
Thank You.