| Bug #30888 | Innodb table + stored procedure + row deletion = server crash | ||
|---|---|---|---|
| Submitted: | 7 Sep 2007 11:04 | Modified: | 8 Oct 2007 18:23 |
| Reporter: | Alexey Botchkov | ||
| Status: | Closed | ||
| Category: | Server: InnoDB | Severity: | S3 (Non-critical) |
| Version: | 5.1.22 | OS: | Any |
| Assigned to: | Sunny Bains | Target Version: | |
[7 Sep 2007 11:07]
Alexey Botchkov
test for the bug
Attachment: rpl_innodb_bug.test (application/octet-stream, text), 1.61 KiB.
[7 Sep 2007 19:32]
Omer BarNir
while verifying hf's patch to this bug observed bug=30907 exposing an auto_inc problem with innodb partitioned tables (not related to the patch) when using --innodb_autoinc_lock_mode=0
[8 Sep 2007 17:11]
Heikki Tuuri
Oops, I should have spotted this bug in Sunny's patch. http://dev.mysql.com/doc/refman/5.0/en/numeric-types.html A MEDIUMINT is 3 bytes! Assigning this bug for Sunny to fix.
[8 Sep 2007 18:09]
Heikki Tuuri
Alexey's patch is correct, of course.
[5 Oct 2007 19:56]
Bugs System
Pushed into 5.1.23-beta
[5 Oct 2007 20:09]
Calvin Sun
the fix is already in 5.1.22 release.
[8 Oct 2007 18:23]
Jon Stephens
Thank you for your bug report. This issue has already been fixed in the latest released version of that product, which you can download at http://www.mysql.com/downloads/ Closed without further action per comment from Calvin.
[21 Nov 2007 19:53]
Bugs System
Pushed into 5.1.23-rc
[21 Nov 2007 19:53]
Bugs System
Pushed into 6.0.4-alpha
Description: the #28430 bug was reopened as it's test case failed. But it appeared that the new problem is not about PARTITION engine, but about Innobase (can be reproduced on just InnoDB tables), and was introduced with the last bit Tim's patch to the InnoDB. So i decided to create a new entry for this issue. How to repeat: put the rpl_innodb_bug.test to the mysql-test/t then execute in mysql-test: ./mysql-test-run.pl --mysqld='--innodb' rpl_innodb_bug Suggested fix: The assertion happens when slave handles Delete event. Here is the patch that fixes the problem on my computer. ===== storage/innobase/row/row0sel.c 1.120 vs edited ===== *** /tmp/bk_row0sel.c-1.120_zkOVSy 2007-09-07 12:52:58 +05:00 --- edited/storage/innobase/row/row0sel.c 2007-09-07 12:50:24 +05:00 *************** *** 4587,4592 **** --- 4587,4597 ---- value = *(ib_uint32_t*) ptr; break; + case 3: + value = *(ib_uint32_t*) ptr; + value&= 0xFFFFFF; + break; + case 2: value = *(uint16 *) ptr; break;