The spark/yarn/hive jobs uses local directories for localization purpose. The required data which is used by the application is stored on local directories when the application is in execution.   Please manually delete the older data from local /tmp directory. Also,   Please follow the steps mentioned in following article to clear the cache memory from local directory : https://community.cloudera.com/t5/Community-Articles/How-to-clear-local-file-cache-and-user-cache-for-yarn/ta-p/245160   Please manually delete the unwanted data from local /tmp directory and also follow the above article   If you found that the provided solution(s) assisted you with your query, please take a moment to login and click Accept as Solution below each response that helped.   ... View more

To resolve this issue, open /opt/cloudera/cm/bin/gen_credentials.sh for editing and add the following to the very end of the script: exit ${PIPESTATUS[0]} And then try generating Kerberos credentials once more.  ... View more

@techfriend this can be resolved after modifiying the principle. WARNING: no policy specified for mapred/ip-172-31-46-169.us-west-2.compute.internal@HADM.RU; defaulting to no policy add_principal: Principal or policy already exists while creating "mapred/ip-172-31-46-169.us-west-2.compute.internal@HADM.RU". + '[' 604800 -gt 0 ']' ++ kadmin -k -t /var/run/cloudera-scm-server/cmf5922922234613877041.keytab -p cloudera-scm/admin@HADM.RU -r HADM.RU -q 'getprinc -terse mapred/ip-172-31-46-169.us-west-2.compute.internal@HADM.RU' ++ tail -1 ++ cut -f 12 + RENEW_LIFETIME=0 + '[' 0 -eq 0 ']' + echo 'Unable to set maxrenewlife' + exit 1 Login to kadmin.local shell then modify the principle using below comamnd. kadmin.local modprinc -maxrenewlife 90day +allow_renewable mapred/ip-172-31-46-169.us-west-2.compute.internal@HADM.RU   ... View more