As seen below, it looks like we're doing a few Kerberos-enabling steps without first checking to see whether the cluster even has Kerberos enabled.
|
# Following is needed for Kerberos and Kafka to work correctly. |
|
logger.info('Copying streamsets keytab to a fixed location which is shared on all clustered nodes ...') |
|
commands = [('cp "$(find /var/run/cloudera-scm-agent/process/*streamsets-DATACOLLECTOR -maxdepth 0 -mindepth 0 | ' |
|
'sort -rs | head -1)/streamsets.keytab" {}/streamsets.keytab').format(KERBEROS_CONFIG_CONTAINER_DIR), |
|
'chown sdc:sdc {}/streamsets.keytab'.format(KERBEROS_CONFIG_CONTAINER_DIR), |
|
'chown sdc:sdc {}'.format(JAAS_CONFIG_FILE_PATH)] |
|
cluster.primary_node.execute(' && '.join(commands)) |
As seen below, it looks like we're doing a few Kerberos-enabling steps without first checking to see whether the cluster even has Kerberos enabled.
topology_cdh/start.py
Lines 1276 to 1282 in d89d510