I am currently using a graph that i load from a file when i run my Graphx application locally.
I'd like to run the application in cluster standalone mode.
Do I have to make changes like place the file in each cluster node? Can I leave my application unchanged and just keep the file in the driver?
Thank you.
In order to allow the executors on the node to access an input file, the file should be access by the nodes.
The preferred way is to read the file from a location which support multi nodes, e.g. HDFS, cassandra
It is possible that placing a copy of the file on each node might work as well, but it isn't the recommended way.