Issues in Metastore Migration

[harshfadiya]

Hi Team,

As I'm performing the Databricks workspace migration, during Metastore migration I'm facing below issue.
As we found differences in the Metastore table count between Legacy and Target workspace, we checked error logs.
After going through Failed Export/Import Metastore logs we have categorize errors in below categories.

1. PATH DOESN'T EXIST (AnalysisException)
   2, The Specified Schema does not match with existing schema
2. java.util.concurrent.ExecutionException: java.lang.ClassNotFoundException
3. InvalidMountException
4. ParseException

Please review above points and let me know what could be the Root Cause of this issues and possible resolutions.
That'll help us to proceed further with Metastore Migrations.

Thanks in advance.

failed_export_metastore.log
failed_import_metastore.log

[dmoore247]

@harshfadiya
Broken tables (see your for exception types) will prevent the apis from exporting the table.
Broken tables are often caused by
an expired or non-existent credential,
the storage location has been moved or modified or removed.
Mounts could be removed or have an expired credential as well which will in turn cause the describe of the table to fail.
Creating a table on an existing Delta Lake table location may fail if the DDL no longer matches the schema defined by the Delta Lake transaction log.

Fix or remove the broken table and you should get a cleaner run of the migration.