Re-classify bulk_imports as gitlab_main_cell with an appropriate sharding key and investigate if other tables have not been classified correctly
It appears that bulk_imports may not be classified correctly for the sharding required for Cells:
Given the FK constraints, we need ensure all related tables have sharding keys. For example
bulk_importsneeds a sharding key (and re-classified asgitlab_main_cell)
- @tkuah in #463854 (comment 2165117931)
We should re-classify this table, set the appropriate sharding keys, and check to see if there are any other tables related to importers that should also be re-classified.
As part of this work, we also need to ensure that a bulk_imports record is not referenced by bulk_import_entities belonging to different organizations.
The bulk_import_configurations also needs a sharding key set, based on what is decided with the bulk_imports. So this should be updated as part of this issue as well.
Implementation Plan
- Create an
organization_idforbulk_imports - When creating new
BulkImportfrom the API, validate that all theBulkImports::Entitybelong to the same organization - Update
db/docs/bulk_imports.ymlanddb/docs/bulk_import_configurations.ymlwith the correct sharding keys