In the unzipped file
structure, you must edit the file Admin/etc/sas_hadoop_config.properties to
supply certain information that cannot be obtained automatically.
Optional settings also exist that you might want to enable.
For the following section:
hadoop.client.config.filepath=<replace with full path>/User/SASWorkspace/hadoop/conf
hadoop.client.jar.filepath=<replace with full path>/User/SASWorkspace/hadoop/lib
hadoop.client.repository.path=<replace with full path>/User/SASWorkspace/hadoop/repository/
hadoop.client.configfile.repository=<replace with full path>/User/SASWorkspace/hadoop/repository
Replace <replace
with full path> with the full path to the location
where the ZIP file was unzipped.
For the following section:
hadoop.cluster.manager.hostname=
hadoop.cluster.manager.port=
hadoop.cluster.hivenode.admin.account=
hadoop.cluster.manager.admin.account=
Set hadoop.cluster.manager.hostname to
the value of the host where either Cloudera Manager or Ambari is running.
Set hadoop.cluster.manager.port to
the value of the port on which Cloudera Manager or Ambari is listening.
Default values are provided.
Set hadoop.cluster.hivenode.admin.account to
the value of a valid account on the machine on which the Hive2 service
is running.
Set hadoop.cluster.manager.admin.account to
the value of a valid Cloudera Manager or Ambari account.
For the following section:
hadoop.client.sasconfig.logfile.path=logs
hadoop.client.sasconfig.logfile.name=logs/sashadoopconfig/sashadoopconfig.log
hadoop.client.config.log.level=0
The default values of logs and sashadoopconfig.log create
the directory Admin/logs and the filename sashadoopconfig.log,
respectively. Both of these values can be changed if you prefer.
You can set the value
of hadoop.client.config.log.level to 3 to
increase the amount of information logged.
Note: If your distribution is secured
with Kerberos,
-
set hadoop.cluster.hivenode.credential.type=kerberos
-
set hadoop.client.config.log.level=3
If you use Cloudera
Manager and it manages multiple clusters, provide the name of the
cluster to use for the value of hadoop.cluster.manager.clustername=.