Is Knime Spark executor supported on Mapr 5.2 spark?
it's not in the list:
KNIME Spark Executor | KNIME
Hi denny marton,
my name is Björn Lohrmann, I work for KNIME on KNIME Spark Executor.
The most important aspect of integrating KNIME Spark Executor with MapR is the the installation of Spark Jobserver on the MapR cluster. Spark Jobserver provides a REST interface to submit Spark jobs from the KNIME client to the cluster.
While we have no official support for MapR yet, if you want to use MapR with KNIME, my recommendation is as follows:
Use one of our generic Spark/HDP builds that has a matching Spark version. For example, if you want to use Spark 2.1 on MapR you should use the build for HDP 2.5 (see ). I think the only thing you have to change is to set SPARK_HOME in setenv.sh so that it point to the correct Spark installation directory on MapR. I cannot guarantee that this works out-of-the-box, but I would be willing to provide some support if problems arise. Under  you also find a link to the latest installation guide PDF that describes the steps of installing Spark Jobserver.
If you have any further questions we can also discuss them in the KNIME Big Data Extensions forum .
 KNIME Spark Executor | KNIME
 Big Data Extensions | KNIME (forum login required)
Hello denny marton. Just for the sake of clarity - are you looking for a high level compatibility matrix between Actian KNIME and MapR 5.2? Or, are you specifically looking for answers surrounding the KNIME Spark executor compatibility with MapR 5.2?
If it is the former, then going by these resources 1, 2, etc. I'd assume yes. But if it is the latter, then it'd help to know the specifics of the issue you are faced with like,nv
=> Something that worked previously is currently found to be broken.=> Error details.=> Etc.
I'm inviting Vinayak Meghraj also into this thread to gather his inputs on this as well.
I did not looking for anything related to Actian.
We planned using KNIME to ease development of machine learning model, thinking about deploy / run it on SPARK on mapr.
One of the way for KNIME to interface/execute flow on SPARK is to use Spark executor component, but upon checking the compatibility list I saw only cloudera/Hortonworks is supported.
Was wondering can I use Spark on Mapr for it?
I have not actually tried it, just checking whether its compatible or not. / anyone have experience on it.
I am inviting John Ewing to share his experience using KNIME. Thank you.
yes, thanks Bjorn, I actually got the same recommendations after writing to KNIME support, they are very responsive in answering my queries.
In the end I end up using the KNIME executors for HDP 2.6, this is the most similar spark version 2.1, just minor adjustment needed on the configuration steps. Everything seems to works ok, can build a job and use spark module and run on MapR spark.
Anyone curious whether KNIME spark executor would run on MapR, I can attest that it can.
Thank you so much for sharing your progress and testimony with the community.
Retrieving data ...