def _prepare_kerberos(cluster, instances=None):
if instances is None:
kerberos.deploy_infrastructure(cluster, get_kdc_server(cluster))
kerberos.prepare_policy_files(cluster)
else:
server = None
if not kerberos.using_existing_kdc(cluster):
server = get_kdc_server(cluster)
kerberos.setup_clients(cluster, server)
kerberos.prepare_policy_files(cluster)
def prepare_scaling_kerberized_cluster(cluster, cloudera_utils, instances):
if kerberos.is_kerberos_security_enabled(cluster):
server = None
if not kerberos.using_existing_kdc(cluster):
server = cloudera_utils.pu.get_manager(cluster)
kerberos.setup_clients(cluster, server)
kerberos.prepare_policy_files(cluster)
# manager can correctly handle updating configs
cloudera_utils.push_kerberos_configs(cluster)
kerberos.create_keytabs_for_map(
cluster,
{'hdfs': cloudera_utils.pu.get_hdfs_nodes(cluster, instances)})
def setup_kerberos_for_cluster(cluster, cloudera_utils):
if kerberos.is_kerberos_security_enabled(cluster):
manager = cloudera_utils.pu.get_manager(cluster)
kerberos.deploy_infrastructure(cluster, manager)
cloudera_utils.full_cluster_stop(cluster)
kerberos.prepare_policy_files(cluster)
cloudera_utils.push_kerberos_configs(cluster)
cloudera_utils.full_cluster_start(cluster)
kerberos.create_keytabs_for_map(
cluster, {
'hdfs': cloudera_utils.pu.get_hdfs_nodes(cluster),
'spark': [cloudera_utils.pu.get_spark_historyserver(cluster)]
})