[ https://issues.apache.org/jira/browse/EAGLE-54?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15025955#comment-15025955 ]
Hao Chen commented on EAGLE-54: ------------------------------- It can be easily done by rewriting this line: https://github.com/apache/incubator-eagle/blob/bf35ede03a55df9c851669e16c59049f696e0b6d/eagle-security/eagle-security-hdfs-auditlog/src/main/java/org/apache/eagle/security/auditlog/HdfsAuditLogKafkaDeserializer.java#L46 My only concern is why not keep using the consistent native Kafka deserializer instead of customized Eagle deserializer, which may include another interface for user to care about. > Encoding match between Kafka serialization and Eagle deserialization > -------------------------------------------------------------------- > > Key: EAGLE-54 > URL: https://issues.apache.org/jira/browse/EAGLE-54 > Project: Eagle > Issue Type: Improvement > Reporter: Edward Zhang > Priority: Minor > > For various audit logs, Kafka serializes log with String type into bytes, and > Eagle deserializes bytes into String. Kafka supports to configure > serialization class by specifying kafka.serializer.StringEncoder and encoding > type. But Eagle lack the mechanism to deserialize with configuration of > encoding, today it does not have issue as it by default uses UTF-8 both. But > it is more graceful for Eagle to expose encoding configuration to be aligned > with Kafka. > https://github.com/apache/kafka/blob/b62f8ea43b6d5307f7274fbe8b7984dd5ee22239/core/src/main/scala/kafka/serializer/Encoder.scala > https://github.com/apache/incubator-eagle/blob/bf35ede03a55df9c851669e16c59049f696e0b6d/eagle-security/eagle-security-hdfs-auditlog/src/main/java/org/apache/eagle/security/auditlog/HdfsAuditLogKafkaDeserializer.java -- This message was sent by Atlassian JIRA (v6.3.4#6332)