Madhan Neethiraj created RANGER-5151:
----------------------------------------

             Summary: Audit to HDFS fails with error: Stream encountered errors 
while writing audits to HDFS!
                 Key: RANGER-5151
                 URL: https://issues.apache.org/jira/browse/RANGER-5151
             Project: Ranger
          Issue Type: Bug
          Components: audit
    Affects Versions: 2.6.0, 2.5.0, 2.4.0, 2.3.0
            Reporter: Madhan Neethiraj


RANGER-3603 introduced an option to periodically check for rollover of the HDFS 
file currently open to write audit logs. This check closes the file if it finds 
that the time for rollover is already past. However, this can result in 
following error if another thread is in the middle of writting to the file:
{noformat}
2025-02-19T01:54:57,711  INFO [AuditFilePeriodicRollOverTask] 
utils.AbstractRangerAuditWriter: Closing file. Rolling over. name = hdfs, 
fileName = 
hdfs://ranger-hadoop:9000/ranger/audit/hiveServer2/20250219/hiveServer2_ranger_audit_ranger-hive.example.com.6.log
2025-02-19T01:54:58,116 ERROR [org.apache.ranger.audit.queue.AuditBatchQueue1] 
utils.RangerJSONAuditWriter: Stream encountered errors while writing audits to 
HDFS!
2025-02-19T01:54:58,116  WARN [org.apache.ranger.audit.queue.AuditBatchQueue1] 
provider.BaseAuditHandler: failed to log audit event: 
{"repoType":3,"repo":"dev_hive","reqUser":"madhan","evtTime":"2025-02-19 
01:54:54.981","access":"SELECT","resource":"test_madhan/test_tbl1/dob,id,name","resType":"@column","action":"select","result":1,"agent":"hiveServer2","policy":7,"reason":null,"enforcer":"ranger-acl","sess":"adbab684-f28a-4f75-bc65-9dd213b7fa58","cliType":"HIVESERVER2","cliIP":"172.18.0.13","reqData":"select
 * from 
test_madhan.test_tbl1","agentHost":"ranger-hive.example.com","logType":"RangerAudit","id":"22008707-e40a-4f37-8741-767703aee89f-0","seq_num":1,"event_count":1,"event_dur_ms":0,"tags":[],"datasets":null,"projects":null,"additional_info":"{\"forwarded-ip-addresses\":\"[]\",\"remote-ip-address\":\"172.18.0.13\"}","cluster_name":"","zone_name":null,"policy_version":1}
 , errorMessage=
{noformat}
 
To prevent this error, closing of the file must be performed in the same thread 
that writes to the file.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to