[ https://issues.apache.org/jira/browse/ATLAS-3902?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ashutosh Mestry updated ATLAS-3902: ----------------------------------- Description: *Background* Consider the scenario where 2 clusters containing Atlas are setup to be synchronized Atlas' export and import APIs. If the source Atlas has changes where table is dropped and re-created with same name. The table's entity within Atlas will get a new GUID but will continue to have the same _qualifiedName_. This case is handled within the Import API. However, the case that is not handled is to perform similar update on to the table's storage descriptor. *Steps to Duplicate* # Create a schema within Hive containing database, tables, columns and views. Atlas will reflect the changes. Perform export. Generate _s1.zip_. # Drop schema. # Re-create the same schema within Hive. Perform export. Generate _s2.zip_. # Clear Atlas database. # Import _s1.zip_. Observe _application.log_. # Import s2.zip. Observe _application.log_. During import log will generate messages like '_GUID Updated: Entity..._' _Expected result:_ Import should succeed with messages indicating changes entity's GUID. _Actual result_: Import fails with errors indicating schema violation (_AtlasSchemaViolation_) was: *Background* *Steps to Duplicate* # Create a schema within Hive containing database, tables, columns and views. Atlas will reflect the changes. Perform export. Generate _s1.zip_. # Drop schema. # Re-create the same schema within Hive. Perform export. Generate _s2.zip_. # Clear Atlas database. # Import _s1.zip_. Observe _application.log_. # Import s2.zip. Observe _application.log_. During import log will generate messages like '_GUID Updated: Entity..._' _Expected result:_ Import should succeed with messages indicating changes entity's GUID. _Actual result_: Import fails with errors indicating schema violation (_AtlasSchemaViolation_) > Import Service: Importing Data With Differing GUIDs for Same Unique > Attributes Causes Errors > -------------------------------------------------------------------------------------------- > > Key: ATLAS-3902 > URL: https://issues.apache.org/jira/browse/ATLAS-3902 > Project: Atlas > Issue Type: Bug > Components: atlas-core > Affects Versions: 2.0.0, trunk, 2.1.0 > Reporter: Ashutosh Mestry > Assignee: Ashutosh Mestry > Priority: Major > Fix For: trunk, 2.1.0 > > > *Background* > Consider the scenario where 2 clusters containing Atlas are setup to be > synchronized Atlas' export and import APIs. If the source Atlas has changes > where table is dropped and re-created with same name. The table's entity > within Atlas will get a new GUID but will continue to have the same > _qualifiedName_. > This case is handled within the Import API. > However, the case that is not handled is to perform similar update on to the > table's storage descriptor. > *Steps to Duplicate* > # Create a schema within Hive containing database, tables, columns and > views. Atlas will reflect the changes. Perform export. Generate _s1.zip_. > # Drop schema. > # Re-create the same schema within Hive. Perform export. Generate _s2.zip_. > # Clear Atlas database. > # Import _s1.zip_. Observe _application.log_. > # Import s2.zip. Observe _application.log_. During import log will generate > messages like '_GUID Updated: Entity..._' > _Expected result:_ Import should succeed with messages indicating changes > entity's GUID. > _Actual result_: Import fails with errors indicating schema violation > (_AtlasSchemaViolation_) -- This message was sent by Atlassian Jira (v8.3.4#803005)