When a new is created, MyRocks assigns new index id "current maximum index id + 1". Current maximum index id is currently not persisted into data dictionary, but calculated by scanning two data dictionaries -- Table Name => internal index id mappings and Ongoing drop index entry. This may introduce a bug of assigning duplicate index, especially when introducing new data dictionary format in the future.
This task is to create another data dictionary to store and update maximum index id. The id should be referenced and updated when new index is created.
When a new is created, MyRocks assigns new index id "current maximum index id + 1". Current maximum index id is currently not persisted into data dictionary, but calculated by scanning two data dictionaries -- Table Name => internal index id mappings and Ongoing drop index entry. This may introduce a bug of assigning duplicate index, especially when introducing new data dictionary format in the future.
This task is to create another data dictionary to store and update maximum index id. The id should be referenced and updated when new index is created.