Details
-
Type:
Bug
-
Status: Closed
-
Priority:
Minor
-
Resolution: Fixed
-
Affects Version/s: 5.5.32
-
Fix Version/s: 5.5.37
-
Component/s: Storage Engine - InnoDB
-
Labels:
Description
It's an upstream bug (Oracle, Percona), but I keep running into it and jumping into investigating, so I'll file it here to be able to find it faster next time
130825 21:17:59 InnoDB: Assertion failure in thread 140072421771008 in file ha_innodb.cc line 6689 InnoDB: Failing assertion: prebuilt->trx->state == 1
#5 0x00007f6525a2bb8b in __GI_abort () at abort.c:91 #6 0x0000000000ac3efa in ha_innobase::unlock_row (this=0x7f64ac0a1a68) at 5.5/storage/xtradb/handler/ha_innodb.cc:6689 #7 0x00000000007c6af1 in find_all_keys (param=0x7f6526f11e90, select=0x7f649c017c28, sort_keys=0x7f649c0190b0, sort_keys_buf=0x7f649c01afb0 "\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245\245"..., buffpek_pointers=0x7f6526f12190, tempfile=0x7f6526f12000) at 5.5/sql/filesort.cc:656 #8 0x00000000007c5559 in filesort (thd=0x3d33a80, table=0x7f64ac09a810, sortorder=0x7f649c011408, s_length=2, select=0x7f649c017c28, max_rows=18446744073709551615, sort_positions=true, examined_rows=0x7f6526f12498) at 5.5/sql/filesort.cc:240 #9 0x0000000000916cf9 in mysql_delete (thd=0x3d33a80, table_list=0x7f649c005338, conds=0x7f649c00fd20, order_list=0x3d37698, limit=1, options=0) at 5.5/sql/sql_delete.cc:269 #10 0x0000000000611f92 in mysql_execute_command (thd=0x3d33a80) at 5.5/sql/sql_parse.cc:3076 #11 0x0000000000619990 in mysql_parse (thd=0x3d33a80, rawbuf=0x7f649c0051c8 "DELETE FROM `t2` WHERE `i` IN ( SELECT `i` FROM `t1` WHERE `i` = 3 ) ORDER BY `i`, `j` LIMIT 1", length=94, parser_state=0x7f6526f13550) at 5.5/sql/sql_parse.cc:5769 #12 0x000000000060d009 in dispatch_command (command=COM_QUERY, thd=0x3d33a80, packet=0x3d39ee1 "DELETE FROM `t2` WHERE `i` IN ( SELECT `i` FROM `t1` WHERE `i` = 3 ) ORDER BY `i`, `j` LIMIT 1", packet_length=94) at 5.5/sql/sql_parse.cc:1071 #13 0x000000000060c219 in do_command (thd=0x3d33a80) at 5.5/sql/sql_parse.cc:794 #14 0x0000000000713545 in do_handle_one_connection (thd_arg=0x3d33a80) at 5.5/sql/sql_connect.cc:1266 #15 0x0000000000712f2c in handle_one_connection (arg=0x3d33a80) at 5.5/sql/sql_connect.cc:1181 #16 0x00007f65267f1e9a in start_thread (arg=0x7f6526f14700) at pthread_create.c:308 #17 0x00007f6525ae5cbd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112
revision-id: igor@askmonty.org-20130823153435-h0x3db9sjxsov8uc revno: 3860 branch-nick: 5.5
RQG grammar:
query_init: CREATE TABLE IF NOT EXISTS `t1` (`i` INT, `j` INT, KEY(`i`)) ENGINE=InnoDB ; CREATE TABLE IF NOT EXISTS `t2` (`i` INT, `j` INT, KEY(`i`)) ENGINE=InnoDB ; query: INSERT INTO _table ( `i`, `j` ) VALUES ( _digit , _digit ) , ( _digit , _digit ) | DELETE FROM _table WHERE `i` IN ( SELECT `i` FROM _table WHERE `i` = _digit ) ORDER BY `i`, `j` LIMIT 1; _table: `t1` | `t2` ;
RQG command lines:
./gentest.pl --threads=8 --duration=1200 --queries=100M --grammar=1.yy --dsn=dbi:mysql:host=127.0.0.1:port=9306:user=root:database=test
perl ./runall-new.pl --threads=8 --duration=1200 --queries=100M --grammar=1.yy --skip-gendata --basedir=<basedir> --vardir=<vardir>
Gliffy Diagrams
Attachments
Issue Links
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
The upstream bugs are said to have been fixed in 5.5.37; I can't reproduce the problem even on MariaDB 5.5.36 (easily reproducible on 5.5.35). Thus, closing as fixed.