Hi everyone, Thank you for the replies.
@Cos: I am working on a Juju charm that will deploy Apache Bigtop Spark. It seems reasonable that such charm would deploy only Spark by default. When/if you relate the Spark charm to HDFS (or any other supported storage eg Cassandra, HBase etc) it would reconfigure Spark accordingly. @Roman: Indeed these hard coded URLs in the Spark module is a good starting point. Thank you for pointing them out. Thanks, Konstantinos On Sun, May 15, 2016 at 3:10 AM, Roman Shaposhnik <[email protected]> wrote: > On Thu, May 12, 2016 at 3:21 AM, Konstantinos Tsakalozos > <[email protected]> wrote: > > Hi everyone, > > > > Is there a way to deploy spark using the puppet scripts with a storage > layer > > other than HDFS? > > IIRC, the biggest issue with our Puppet code for spark could be hard-coded > URLs for HDFS. E.g.: hdfs://.... > > It'd be a good exercise to make sure those are tweakable (with the default > still > being HDFS). > > Thanks, > Roman. > -- Konstantinos Tsakalozos
