Bug #68311 InnoDB: Failing assertion thr in row0upd.cc on concurrent ALTER, UPDATE, HANDLER

Submitted: 7 Feb 2013 23:21 Modified: 8 Feb 2013 22:02
Reporter: Elena Stepanova Email Updates:
Status: Duplicate Impact on me:
None 
Category:MySQL Server: InnoDB storage engine Severity:S3 (Non-critical)
Version:5.6 OS:Linux (Ubuntu 12.04 64-bit)
Assigned to: CPU Architecture:Any

[7 Feb 2013 23:21] Elena Stepanova
Description:
Version: '5.6.10-debug-log'  
/tmp/master.sock'  port: 19300  Source distribution
2013-02-08 02:08:19 7f3a65e34700  InnoDB: Assertion failure in thread 139888794224384 in file row0upd.cc line 1311
InnoDB: Failing assertion: thr
InnoDB: We intentionally generate a memory trap.

#5  0x00007f3a7b305b8b in __GI_abort () at abort.c:91
#6  0x0000000000c9bd67 in row_upd_changes_ord_field_binary_func (index=0x7f3a30041118, update=0x7f3a30046970, thr=0x0, row=0x0, ext=0x0) at mysql-5.6/storage/innobase/row/row0upd.cc:1311
#7  0x0000000000d17409 in btr_cur_update_in_place (flags=31, cursor=0x7f3a65e2efa0, offsets=0x7f3a3004a760, update=0x7f3a30046970, cmpl_info=0, thr=0x0, trx_id=0, mtr=0x7f3a65e2f0f0) at mysql-5.6/storage/innobase/btr/btr0cur.cc:2000
#8  0x0000000000d179a7 in btr_cur_optimistic_update (flags=31, cursor=0x7f3a65e2efa0, offsets=0x7f3a65e2f088, heap=0x7f3a65e2ef70, update=0x7f3a30046970, cmpl_info=0, thr=0x0, trx_id=0, mtr=0x7f3a65e2f0f0) at mysql-5.6/storage/innobase/btr/btr0cur.cc:2122
#9  0x0000000000d18334 in btr_cur_pessimistic_update (flags=31, cursor=0x7f3a65e2efa0, offsets=0x7f3a65e2f088, offsets_heap=0x7f3a65e2ef70, entry_heap=0x7f3a300465e0, big_rec=0x7f3a65e2f0a0, update=0x7f3a30046970, cmpl_info=0, thr=0x0, trx_id=0, mtr=0x7f3a65e2f0f0) at mysql-5.6/storage/innobase/btr/btr0cur.cc:2380
#10 0x0000000000c7277f in row_log_table_apply_update (thr=0x7f3a3003f120, trx_id_col=1, new_trx_id_col=1, mrec=0x7f3a65cb45ab "\200", offsets=0x7f3a30038290, offsets_heap=0x7f3a3004a660, heap=0x7f3a300465e0, dup=0x7f3a65e2f7f0, trx_id=1608, old_pk=0x7f3a30046658) at mysql-5.6/storage/innobase/row/row0log.cc:1688
#11 0x0000000000c73541 in row_log_table_apply_op (thr=0x7f3a3003f120, trx_id_col=1, new_trx_id_col=1, dup=0x7f3a65e2f7f0, error=0x7f3a65e2f7a8, offsets_heap=0x7f3a3004a660, heap=0x7f3a300465e0, mrec=0x7f3a65cb45ab "\200", mrec_end=0x7f3a65cb84e4 "", offsets=0x7f3a30038290) at mysql-5.6/storage/innobase/row/row0log.cc:2033
#12 0x0000000000c745a5 in row_log_table_apply_ops (thr=0x7f3a3003f120, dup=0x7f3a65e2f7f0) at mysql-5.6/storage/innobase/row/row0log.cc:2326
#13 0x0000000000c749e8 in row_log_table_apply (thr=0x7f3a3003f120, old_table=0x7f3a4403ba28, table=0x7f3a300365f0) at mysql-5.6/storage/innobase/row/row0log.cc:2420
#14 0x0000000000b90436 in ha_innobase::inplace_alter_table (this=0x7f3a30008930, altered_table=0x7f3a300365f0, ha_alter_info=0x7f3a65e2fa50) at mysql-5.6/storage/innobase/handler/handler0alter.cc:3842
#15 0x000000000083e45d in handler::ha_inplace_alter_table (this=0x7f3a30008930, altered_table=0x7f3a300365f0, ha_alter_info=0x7f3a65e2fa50) at mysql-5.6/sql/handler.h:2828
#16 0x0000000000837131 in mysql_inplace_alter_table (thd=0x3c2f8a0, table_list=0x7f3a30004fa8, table=0x7f3a30008050, altered_table=0x7f3a300365f0, ha_alter_info=0x7f3a65e2fa50, inplace_supported=HA_ALTER_INPLACE_NO_LOCK_AFTER_PREPARE, target_mdl_request=0x7f3a65e31170, alter_ctx=0x7f3a65e2fb50) at mysql-5.6/sql/sql_table.cc:6419
#17 0x000000000083b0ef in mysql_alter_table (thd=0x3c2f8a0, new_db=0x7f3a30005500 "test", new_name=0x0, create_info=0x7f3a65e313b0, table_list=0x7f3a30004fa8, alter_info=0x7f3a65e31490, order_num=0, order=0x0, ignore=false) at mysql-5.6/sql/sql_table.cc:8098
#18 0x000000000097aecf in Sql_cmd_alter_table::execute (this=0x7f3a30005698, thd=0x3c2f8a0) at mysql-5.6/sql/sql_alter.cc:313
#19 0x00000000007ce64e in mysql_execute_command (thd=0x3c2f8a0) at mysql-5.6/sql/sql_parse.cc:4871
#20 0x00000000007d164d in mysql_parse (thd=0x3c2f8a0, rawbuf=0x7f3a30004db0 "ALTER TABLE `table1000_innodb_int_autoinc` ADD COLUMN filler VARCHAR(255) DEFAULT ' filler '", length=92, parser_state=0x7f3a65e33110) at mysql-5.6/sql/sql_parse.cc:6149
#21 0x00000000007c4eda in dispatch_command (command=COM_QUERY, thd=0x3c2f8a0, packet=0x3c335b1 "ALTER TABLE `table1000_innodb_int_autoinc` ADD COLUMN filler VARCHAR(255) DEFAULT ' filler '", packet_length=92) at mysql-5.6/sql/sql_parse.cc:1334
#22 0x00000000007c3fe2 in do_command (thd=0x3c2f8a0) at mysql-5.6/sql/sql_parse.cc:1036
#23 0x000000000078c5a8 in do_handle_one_connection (thd_arg=0x3c2f8a0) at mysql-5.6/sql/sql_connect.cc:969
#24 0x000000000078bfac in handle_one_connection (arg=0x3c2f8a0) at mysql-5.6/sql/sql_connect.cc:885
#25 0x0000000000b3ba86 in pfs_spawn_thread (arg=0x3ab8d30) at mysql-5.6/storage/perfschema/pfs.cc:1853
#26 0x00007f3a7bea4e9a in start_thread (arg=0x7f3a65e34700) at pthread_create.c:308

Trying to get some variables.
Some pointers may be invalid and cause the dump to abort.
Query (7f3a30004db0): ALTER TABLE `table1000_innodb_int_autoinc` ADD COLUMN filler VARCHAR(255) DEFAULT ' filler '
Connection ID (thread ID): 7
Status: NOT_KILLED

revision-id: daniel.fischer@oracle.com-20130201133216-o0wf9m4ns9hpdkfo
revno: 4806
branch-nick: mysql-5.6

How to repeat:
# RQG grammar (bug.yy):

query:
	handler_open_close |
	handler_read ;

thread1:
	UPDATE IGNORE _table SET _field = value where ;

thread2: 
	ALTER TABLE _table ADD COLUMN filler VARCHAR(255) DEFAULT ' filler ' |
	ALTER TABLE _table DROP COLUMN filler ;

handler_open_close:
	HANDLER alias1 CLOSE ; HANDLER _table OPEN AS alias1 ;

handler_read:
	HANDLER alias1 READ `PRIMARY` index_op where ;

comp_op:
	= | <= | >= | < | > ;

index_op:
	FIRST | NEXT | PREV | LAST ;

value:
	_digit | _tinyint_unsigned | _varchar(1) ;

where:
	| WHERE _field comp_op value ;

# End of RQG grammar

#########################

# RQG data template (bug.zz):

$tables = {
        rows => [1000],
	engines => [ 'InnoDB' ]
};

$fields = {
        types => [ 'int', 'char(32)' ],
        indexes => [ 'key', 'unique' ],
        null => [undef, 'not null'],
};

$data = {
        numbers => [ 'digit', 'null', undef , 'tinyint_unsigned' ],
        strings => [ 'letter', 'english', 'varchar(3)' ]
}

# End of RQG data template

##############################

RQG command line:

perl runall.pl --grammar=bug.yy --gendata=bug.zz --basedir=mysql-5.6 --queries=100K --duration=600 --threads=3 --mysqld=--innodb-lock-wait-timeout=1 --mysqld=--lock-wait-timeout=1 

# It fails within a minute after start for me, 
# but if it doesn't for you, try to increase the number of threads
[8 Feb 2013 11:22] Erlend Dahl
I have tried to reproduce the problem, but no luck so far, even with the number of threads increased to 20. 

We will give it a try in another environment.
[8 Feb 2013 11:47] Elena Stepanova
For a note, my server was built with BUILD/compile-pentium-debug-max-no-ndb.

The machine is Ubuntu 12.04 64-bit (as indicated), in a VM with 4 Gb memory and 2 cores. 

The randgen is from the current tree.

If it doesn't work with the provided grammar/data files, you could also try standard conf/engines/handler.yy and conf/engines/handler.zz, correspondingly. They produce a lot of noise due to unsupported syntax and generally take me longer to reproduce the problem, but maybe it will be different for you.
[8 Feb 2013 12:55] Erlend Dahl
Thanks, that was very useful. I can reproduce on 5.6.10 but not on recent 5.6.

I will have somebody take a closer look to see what the root cause can be.
[8 Feb 2013 14:22] MySQL Verification Team
On Windows I couldn't repeat with 5.6.10 source package with the original test and using conf/engines/handler.yy and conf/engines/handler.zz. Windows OS is Windows 7 64-bit 8GB machine.
[8 Feb 2013 22:02] Calvin Sun
it is a duplicate of an internal bug, which is fixed (will be in 5.6.11).
[8 Feb 2013 22:02] Calvin Sun
Internal bug#16063835

Portions of this website are copyright © 2001, 2002 The PHP Group

Page generated in 0.042 sec. using MySQL 8.4.7-u6-cloud

Timestamp references displayed by the system are UTC.
Content reproduced on this site is the property of the respective copyright holders.
It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.