New Contributor
Posts: 2
Registered: ‎08-09-2017

Cloudera cluster with out HDFS and NFS gateway. Only using NFS

Hi Big data geeks
I am planning to install Cloudera cluster without using HDFS and NFS Gateway. I will be only using NFS location, is this possible. I am ok to have some meta data and config on HDFS but all the data should be stored in NFS location. Is this possible?
I have installed right now with CM and I am trying to change location to NFS mount point location but CM thinks it is HDFS location and creating it under HDFS only.
Full points for any answers.
Thanks and best regards

Posts: 1,892
Kudos: 432
Solutions: 302
Registered: ‎07-31-2013

Re: Cloudera cluster with out HDFS and NFS gateway. Only using NFS

Since its not super-clear from the description text, I am assuming you have an storage solution mountable via NFS, and you'd like to use this storage directly instead of running HDFS on disks.

While it may work to some degree, with the file:/// URIs pointing to the NFS location as its root in all relevant path configurations, I'd strongly recommend against this as NFS is not designed for the style of workload that YARN/etc. run, and there'll be numerous consistency problems too.

Have you thought of using alternative, but proper distributed storage systems? S3 for example has good integration with CDH.