Now that the SDK is installed and the routing reference data is configured the reference
data must be distributed around the cluster. Your cluster must be set up for S3 access.
Use the s3a URI scheme on Hortonworks distributions.
For the purpose of this guide, we will install the reference data into
s3a://<your-bucket>/precisely/routing/data.
-
Upload the reference data into S3.
hadoop fs -mkdir -p s3a://<your-bucket>/precisely/routing
hadoop fs -copyFromLocal /precisely/routing/data s3a://<your-bucket>/precisely/routing/data
-
When the data node performs routing tasks, the node will download the reference data
from S3 and onto the local file system. This means a local directory needs to be set up on
all data nodes. Perform the following commands on all data nodes and HiverServer
nodes.
sudo mkdir /precisely/downloads
sudo chown sdkuser:hadoop /precisely/downloads
sudo chmod 775 /precisely/downloads