| Bug #18384 | InnoDB memory leak on duplicate key errors in 5.0 if row has many columns | ||
|---|---|---|---|
| Submitted: | 21 Mar 2006 9:14 | Modified: | 11 Apr 2006 4:15 |
| Reporter: | Heikki Tuuri | ||
| Status: | Closed | ||
| Category: | Server: InnoDB | Severity: | S2 (Serious) |
| Version: | 5.0.3 and later | OS: | Any (All) |
| Assigned to: | Marko Mäkelä | Target Version: | |
[21 Mar 2006 10:14]
Marko Mäkelä
I have committed the patch to the InnoDB source code repository.
[5 Apr 2006 21:19]
Elliot Murphy
Fixed in InnoDB snapshot368; fixes are in 5.0.20.
[11 Apr 2006 4:15]
Paul DuBois
Noted in 5.0.20 changelog. <literal>InnoDB</literal> had a memory leak for duplicate-key errors with tables having 90 columns or more. (Bug #18384)
Description: InnoDB leaks memory in the code below because it forgets to free heap. The leak only happens with rows with about 90 columns or more. row0ins.c: if (cursor->low_match >= n_unique) { rec = btr_cur_get_rec(cursor); if (!page_rec_is_infimum(rec)) { offsets = rec_get_offsets(rec, cursor->index, offsets, ULINT_UNDEFINED, &heap); Marko has a patch here: http://bugs.mysql.com/bug.php?id=17674 How to repeat: Use INSERT IGNORE to a table with many duplicates. Suggested fix: Marko has a patch.