kobotoolbox / kpi / 29867541075

81%

master: 76%

Build Type

push

github

Committed by web-flow

Commit Message

fix(logger): resolve `root_uuid` conflicts in dedup command DEV-2450 (#7292)

### 📣 Summary

A background data-repair task could fail on forms that contain certain
duplicate submissions, leaving those forms unrepaired. It now resolves
the conflict and continues.

### 💭 Notes

The `clean_duplicated_submissions` management command (run by
long-running migration 0027) could raise `unique_root_uuid_per_xform`
while assigning `root_uuid`, which aborted the whole form and left it
flagged as failed.

Two root causes are fixed:
- `_replace_duplicates` gave a divergent duplicate a new unique `uuid`
but left its `root_uuid` pointing at the shared value, because
`_populate_root_uuid()` is a no-op once `root_uuid` is already set. It
now sets `root_uuid` to the new uuid and updates the XML `rootUuid` node
when that node exists.
- `_delete_duplicates` backfilled the reference `root_uuid` without
handling a collision with another submission in the same form.

Both now fall back to a shared helper, `resolve_root_uuid_conflict()` in
`logger.utils.instance`, which assigns a unique `CONFLICT-<uuid4>-...`
`root_uuid` (hashed when it would exceed the column length, and never
using the sequential pk) and keeps Postgres, the XML `rootUuid` node
(when present) and the Mongo `meta/rootUuid` in sync.
`_resolve_conflict_by_pk` in `clean_duplicated_submissions_root_uuid`
now uses the same helper instead of duplicating the logic.

The conflicting `.update()` runs in its own `kc_transaction_atomic`
block so a caught `IntegrityError` cannot poison an enclosing
transaction: it is a savepoint when the command runs inside a
transaction (for example under `TestCase`, which is what surfaced this),
or its own transaction when run standalone in autocommit. Without it,
the conflict resolution below would run on an aborted transaction. In
`_replace_duplicates`, only the identity fields change, so after the
Postgres bulk update Mongo is synced with a single bulk `$set` on
`_uuid`, `meta/inst... (continued)

Coverage Stats

9309 of 12656 branches covered (73.55%)

19 of 19 new or added lines in 2 files covered. (100.0%)

1007 existing lines in 55 files now uncovered.

30343 of 37236 relevant lines covered (81.49%)

5.51 hits per line