[MDEV-22939] Server crashes in row_make_new_pathname | ALTER TABLE ENGINE INNODB after DISCARD TABLESPACE Created: 2020-06-19  Updated: 2020-10-06  Resolved: 2020-09-22

Status: Closed
Project: MariaDB Server
Component/s: Data Definition - Alter Table, Storage Engine - InnoDB
Affects Version/s: 10.3, 10.4
Fix Version/s: 10.2.35, 10.3.26, 10.4.16, 10.5.7

Type: Bug Priority: Critical
Reporter: Shihao Chen Assignee: Marko Mäkelä
Resolution: Fixed Votes: 0
Labels: corruption, not-10.1, not-10.2, not-10.5

Issue Links:
Relates
relates to MDEV-21983 Sig 11 in row_make_new_pathname | REN... Closed
relates to MDEV-23705 Assertion `table->data_dir_path || !s... Closed

 Description    

CREATE TABLE t (a INT) ENGINE=INNODB;
 
ALTER TABLE t DISCARD TABLESPACE;
 
ALTER TABLE t ENGINE INNODB;

Leads to

Exception thrown at 0x00007FF7F7C69C13 in mysqld.exe: 0xC0000005: Access violation reading location 0x0000000000000068.
 
mysqld.exe!row_make_new_pathname(dict_table_t * table, const char * new_name) Line 4280 (d:\maria\server\storage\innobase\row\row0merge.cc:4280)
 
mysqld.exe!row_merge_rename_tables_dict(dict_table_t * old_table, dict_table_t * new_table, const char * tmp_name, trx_t * trx) Line 4335 (d:\maria\server\storage\innobase\row\row0merge.cc:4335)
 
mysqld.exe!commit_try_rebuild(Alter_inplace_info * ha_alter_info, ha_innobase_inplace_ctx * ctx, TABLE * altered_table, const TABLE * old_table, trx_t * trx, const char * table_name) Line 8508 (d:\maria\server\storage\innobase\handler\handler0alter.cc:8508)
 
mysqld.exe!ha_innobase::commit_inplace_alter_table(TABLE * altered_table, Alter_inplace_info * ha_alter_info, bool commit) Line 9473 (d:\maria\server\storage\innobase\handler\handler0alter.cc:9473)
 
mysqld.exe!handler::ha_commit_inplace_alter_table(TABLE * altered_table, Alter_inplace_info * ha_alter_info, bool commit) Line 4578 (d:\maria\server\sql\handler.cc:4578)
 
mysqld.exe!mysql_inplace_alter_table(THD * thd, TABLE_LIST * table_list, TABLE * table, TABLE * altered_table, Alter_inplace_info * ha_alter_info, enum_alter_inplace_result inplace_supported, MDL_request * target_mdl_request, Alter_table_ctx * alter_ctx) Line 7678 (d:\maria\server\sql\sql_table.cc:7678)
 
mysqld.exe!mysql_alter_table(THD * thd, const st_mysql_const_lex_string * new_db, const st_mysql_const_lex_string * new_name, HA_CREATE_INFO * create_info, TABLE_LIST * table_list, Alter_info * alter_info, unsigned int order_num, st_order * order, bool ignore) Line 9918 (d:\maria\server\sql\sql_table.cc:9918)
 
mysqld.exe!Sql_cmd_alter_table::execute(THD * thd) Line 500 (d:\maria\server\sql\sql_alter.cc:500)
 
mysqld.exe!mysql_execute_command(THD * thd) Line 6030 (d:\maria\server\sql\sql_parse.cc:6030)
 
mysqld.exe!mysql_parse(THD * thd, char * rawbuf, unsigned int length, Parser_state * parser_state, bool is_com_multi, bool is_next_command) Line 7817 (d:\maria\server\sql\sql_parse.cc:7817)
 
mysqld.exe!dispatch_command(enum_server_command command, THD * thd, char * packet, unsigned int packet_length, bool is_com_multi, bool is_next_command) Line 1858 (d:\maria\server\sql\sql_parse.cc:1858)
 
mysqld.exe!do_command(THD * thd) Line 1401 (d:\maria\server\sql\sql_parse.cc:1401)
 
mysqld.exe!threadpool_process_request(THD * thd) Line 359 (d:\maria\server\sql\threadpool_common.cc:359)
 
mysqld.exe!tp_callback(TP_connection * c) Line 186 (d:\maria\server\sql\threadpool_common.cc:186)
 
mysqld.exe!tp_callback(_TP_CALLBACK_INSTANCE * instance, void * context) Line 376 (d:\maria\server\sql\threadpool_win.cc:376)
 
mysqld.exe!work_callback(_TP_CALLBACK_INSTANCE * instance, void * context, _TP_WORK * work) Line 450 (d:\maria\server\sql\threadpool_win.cc:450)
 
ntdll.dll!00007ffc3c03081d() (Unknown Source:0)
 
ntdll.dll!00007ffc3c0345b4() (Unknown Source:0)
 
kernel32.dll!00007ffc3bd67bd4() (Unknown Source:0)
 
ntdll.dll!00007ffc3c06ce51() (Unknown Source:0)

In 10.3, 10.4, 10.5. Not found in 10.2.

Result of above queries in 10.2:

create table t(a INT) ENGINE = INNODB;
 
Query OK, 0 rows affected (0.32 sec)
 
 
 
mysql> ALTER TABLE t DISCARD TABLESPACE;
 
Query OK, 0 rows affected (0.09 sec)
 
 
 
ALTER TABLE t ENGINE INNODB;
 
Query OK, 0 rows affected, 1 warning (0.33 sec)
 
Records: 0  Duplicates: 0  Warnings: 1
 
 
 
show warnings;
 
+---------+------+---------------------------------------------+
 
| Level   | Code | Message                                     |
 
+---------+------+---------------------------------------------+
 
| Warning | 1814 | Tablespace has been discarded for table `t` |
 
+---------+------+---------------------------------------------+
 
 
 
select * from t;
 
ERROR 1814 (HY000): Tablespace has been discarded for table `t`



 Comments   
Comment by Shihao Chen [ 2020-06-19 ]

in row_merge_rename_tables_dict@row0merge.cc:

/* Update SYS_TABLESPACES and SYS_DATAFILES if the old table being
 
renamed is a single-table tablespace, which must be implicitly
 
renamed along with the table. */
 
if (err == DB_SUCCESS
 
    && old_table->space_id) {
 
	/* Make pathname to update SYS_DATAFILES. */
 
	char* tmp_path = row_make_new_pathname(old_table, tmp_name);

old_table->space_id and old_table->space->chain.start->name are used in row_make_new_pathname but only old_table->space_id is checked.
Discarded tables seem to have non-zero space_id and NULL space. Are we missing a NULL check here?

Comment by Shihao Chen [ 2020-06-19 ]

Similarly, in fil_mtr_rename_log@fil0fil.cc:

if (old_table->space_id) {
 
        /* omitting unrelated code */
 
	const char* old_path = old_table->space->chain.start->name;

member of space struct is accessed without a null check against old_table->space.

Comment by Shihao Chen [ 2020-06-19 ]

Would like to make the change that aligns the behaviour of 10.3 and after with that of 10.2. Do I create a PR tagging this MDEV directly or is there any pre-work to do?

Comment by Shihao Chen [ 2020-06-19 ]

On a closer look the fix might take more work than I imagined.
In commit_try_rebuild@handler0alter.cc:

/* The new table must inherit the flag from the
 
"parent" table. */
 
if (!user_table->space) {
 
	rebuilt_table->file_unreadable = true;
 
	rebuilt_table->flags2 |= DICT_TF2_DISCARDED;
 
}

So when executing "alter table t engine innodb" on a discarded table, the new table will be created with discarded and unreadable flag but with a tablespace.
Consequently in dict_table_rename_in_cache@dict0dict.cc:

if (!table->space) {
 
    // deleting the tablespace
 
}else{
 
    // renaming the tablespace    
}

As the newly altered table has a tablespace(despite flagged as discarded), it will rename its .ibd into t.ibd, which confuses the server on subsequent queries.
Is it expected behaviour to create a tablespace when inplace altering a discarded table? <- I guess I can work out a fix by checking discarded flag along with tablespace existence, just curious if there's any further reason we create a discarded tablespace in the first place.

Comment by Alice Sherepa [ 2020-06-19 ]

The crash happens on 10.3, 10.4, but not on 10.5

10.3 b633b6a9d8278a3e0d9baaf83

 
#4  row_make_new_pathname (table=0x7f3c4808dc90, new_name=0x7f3c480a8c88 "test/#sql-ib22") at /d1/git/10.3/storage/innobase/row/row0merge.cc:4286
 
#5  0x000055796e090a00 in row_merge_rename_tables_dict (old_table=0x7f3c4808dc90, new_table=0x7f3c480baf60, tmp_name=0x7f3c480a8c88 "test/#sql-ib22", trx=0x7f3c9bc011c8) at /d1/git/10.3/storage/innobase/row/row0merge.cc:4341
 
#6  0x000055796df63a58 in commit_try_rebuild (ha_alter_info=0x7f3c9b29f190, ctx=0x7f3c48013cf8, altered_table=0x7f3c480b9ec0, old_table=0x7f3c48036390, trx=0x7f3c9bc011c8, table_name=0x7f3c4808da25 "t") at /d1/git/10.3/storage/innobase/handler/handler0alter.cc:8602
 
#7  0x000055796df5c05e in ha_innobase::commit_inplace_alter_table (this=0x7f3c480a73d8, altered_table=0x7f3c480b9ec0, ha_alter_info=0x7f3c9b29f190, commit=true) at /d1/git/10.3/storage/innobase/handler/handler0alter.cc:9567
 
#8  0x000055796dce7fd8 in handler::ha_commit_inplace_alter_table (this=0x7f3c480a73d8, altered_table=0x7f3c480b9ec0, ha_alter_info=0x7f3c9b29f190, commit=true) at /d1/git/10.3/sql/handler.cc:4596
 
#9  0x000055796daa6aad in mysql_inplace_alter_table (thd=0x7f3c48000d50, table_list=0x7f3c48012b60, table=0x7f3c48036390, altered_table=0x7f3c480b9ec0, ha_alter_info=0x7f3c9b29f190, inplace_supported=HA_ALTER_INPLACE_COPY_NO_LOCK, target_mdl_request=0x7f3c9b29f310, alter_ctx=0x7f3c9b29f8c0) at /d1/git/10.3/sql/sql_table.cc:7701
 
#10 0x000055796daad251 in mysql_alter_table (thd=0x7f3c48000d50, new_db=0x7f3c48005438, new_name=0x7f3c480057f8, create_info=0x7f3c9b2a04b0, table_list=0x7f3c48012b60, alter_info=0x7f3c9b2a03f0, order_num=0, order=0x0, ignore=false) at /d1/git/10.3/sql/sql_table.cc:9941
 
#11 0x000055796db3b0b4 in Sql_cmd_alter_table::execute (this=0x7f3c480131c8, thd=0x7f3c48000d50) at /d1/git/10.3/sql/sql_alter.cc:512
 
#12 0x000055796d9cce7b in mysql_execute_command (thd=0x7f3c48000d50) at /d1/git/10.3/sql/sql_parse.cc:6030
 
#13 0x000055796d9d25e6 in mysql_parse (thd=0x7f3c48000d50, rawbuf=0x7f3c48012a78 "ALTER TABLE t ENGINE INNODB", length=27, parser_state=0x7f3c9b2a15c0, is_com_multi=false, is_next_command=false) at /d1/git/10.3/sql/sql_parse.cc:7818
 
#14 0x000055796d9bef29 in dispatch_command (command=COM_QUERY, thd=0x7f3c48000d50, packet=0x7f3c48008ed1 "ALTER TABLE t ENGINE INNODB", packet_length=27, is_com_multi=false, is_next_command=false) at /d1/git/10.3/sql/sql_parse.cc:1855
 
#15 0x000055796d9bd84c in do_command (thd=0x7f3c48000d50) at /d1/git/10.3/sql/sql_parse.cc:1401
 
#16 0x000055796db35168 in do_handle_one_connection (connect=0x557970fa7fb0) at /d1/git/10.3/sql/sql_connect.cc:1403
 
#17 0x000055796db34eca in handle_one_connection (arg=0x557970fa7fb0) at /d1/git/10.3/sql/sql_connect.cc:1308
 
#18 0x000055796e4cfd21 in pfs_spawn_thread (arg=0x5579712277e0) at /d1/git/10.3/storage/perfschema/pfs.cc:1869
 
#19 0x00007f3ca6880fa3 in start_thread (arg=<optimized out>) at pthread_create.c:486
 
#20 0x00007f3ca62044cf in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

10.5 605555fc31784302a54d39

 
MariaDB [test]> CREATE TABLE t (a INT) ENGINE=INNODB;
 
Query OK, 0 rows affected (0.017 sec)
 
 
 
MariaDB [test]> ALTER TABLE t DISCARD TABLESPACE;
 
Query OK, 0 rows affected (0.007 sec)
 
 
 
MariaDB [test]> ALTER TABLE t ENGINE INNODB;
 
ERROR 1296 (HY000): Got error 192 '' from InnoDB
 
Warning (Code 1814): Tablespace has been discarded for table `t`
 
Warning (Code 192): Table t in file ./test/t.ibd is encrypted but encryption service or used key_id is not available.  Can't continue reading table.
 
Error (Code 1296): Got error 192 '' from InnoDB

Comment by Roel Van de Paar [ 2020-08-13 ] 

USE test;
 
CREATE TABLE t(c INT) ENGINE=InnoDB;
 
ALTER TABLE t DISCARD TABLESPACE;
 
OPTIMIZE TABLE t;

Leads to:

10.4.15 eae968f62d285de97ed607c87bc131cd863d5d03 (Debug)

 
Core was generated by `/test/MD110820-mariadb-10.4.15-linux-x86_64-dbg/bin/mysqld --no-defaults --core'.
 
Program terminated with signal SIGSEGV, Segmentation fault.
 
#0  __pthread_kill (threadid=<optimized out>, signo=signo@entry=11)
 
    at ../sysdeps/unix/sysv/linux/pthread_kill.c:57
 
[Current thread is 1 (Thread 0x14fc06360700 (LWP 583872))]
 
(gdb) bt
 
#0  __pthread_kill (threadid=<optimized out>, signo=signo@entry=11) at ../sysdeps/unix/sysv/linux/pthread_kill.c:57
 
#1  0x00005623bd6238a6 in my_write_core (sig=sig@entry=11) at /test/10.4_dbg/mysys/stacktrace.c:482
 
#2  0x00005623bcd9fcdc in handle_fatal_signal (sig=11) at /test/10.4_dbg/sql/signal_handler.cc:343
 
#3  <signal handler called>
 
#4  0x00005623bd137fe2 in row_make_new_pathname (table=table@entry=0x14fbdec9a480, new_name=new_name@entry=0x14fbdec42ce8 "test/#sql-ib22") at /test/10.4_dbg/storage/innobase/row/row0merge.cc:4280
 
#5  0x00005623bd13cc12 in row_merge_rename_tables_dict (old_table=old_table@entry=0x14fbdec9a480, new_table=new_table@entry=0x14fbdec9a880, tmp_name=0x14fbdec42ce8 "test/#sql-ib22", trx=trx@entry=0x14fbf0c03380) at /test/10.4_dbg/storage/innobase/row/row0merge.cc:4335
 
#6  0x00005623bd0171b0 in commit_try_rebuild (ha_alter_info=ha_alter_info@entry=0x14fc0635a190, ctx=ctx@entry=0x14fbdec6e630, altered_table=altered_table@entry=0x14fc0635a230, old_table=<optimized out>, trx=<optimized out>, table_name=<optimized out>) at /test/10.4_dbg/storage/innobase/handler/handler0alter.cc:9944
 
#7  0x00005623bd00c462 in ha_innobase::commit_inplace_alter_table (this=0x14fbdec7a088, altered_table=0x14fc0635a230, ha_alter_info=0x14fc0635a190, commit=true) at /test/10.4_dbg/storage/innobase/handler/handler0alter.cc:11016
 
#8  0x00005623bcdadd63 in handler::ha_commit_inplace_alter_table (this=0x14fbdec7a088, altered_table=altered_table@entry=0x14fc0635a230, ha_alter_info=ha_alter_info@entry=0x14fc0635a190, commit=commit@entry=true) at /test/10.4_dbg/sql/handler.cc:4616
 
#9  0x00005623bcba889f in mysql_inplace_alter_table (thd=thd@entry=0x14fbdec15070, table_list=0x14fbdec6d148, table=table@entry=0x14fbdecfd070, altered_table=altered_table@entry=0x14fc0635a230, ha_alter_info=ha_alter_info@entry=0x14fc0635a190, target_mdl_request=target_mdl_request@entry=0x14fc0635b180, alter_ctx=0x14fc0635bcc0) at /test/10.4_dbg/sql/sql_table.cc:7797
 
#10 0x00005623bcbb5c9f in mysql_alter_table (thd=thd@entry=0x14fbdec15070, new_db=<optimized out>, new_name=<optimized out>, create_info=create_info@entry=0x14fc0635c880, table_list=<optimized out>, table_list@entry=0x14fbdec6d148, alter_info=alter_info@entry=0x14fc0635c7c0, order_num=0, order=0x0, ignore=false) at /test/10.4_dbg/sql/sql_table.cc:10156
 
#11 0x00005623bcbb7967 in mysql_recreate_table (thd=thd@entry=0x14fbdec15070, table_list=table_list@entry=0x14fbdec6d148, table_copy=table_copy@entry=false) at /test/10.4_dbg/sql/sql_table.cc:11001
 
#12 0x00005623bcc387f6 in admin_recreate_table (thd=thd@entry=0x14fbdec15070, table_list=table_list@entry=0x14fbdec6d148) at /test/10.4_dbg/sql/sql_admin.cc:59
 
#13 0x00005623bcc3b3f8 in mysql_admin_table (thd=thd@entry=0x14fbdec15070, tables=tables@entry=0x14fbdec6d148, check_opt=check_opt@entry=0x14fbdec1a248, operator_name=operator_name@entry=0x5623bd72a19e "optimize", lock_type=lock_type@entry=TL_WRITE, org_open_for_modify=org_open_for_modify@entry=true, repair_table_use_frm=false, extra_open_options=0, prepare_func=0x0, operator_func=(int (handler::*)(handler * const, THD *, HA_CHECK_OPT *)) 0x5623bcdad856 <handler::ha_optimize(THD*, st_ha_check_opt*)>, view_operator_func=0x0) at /test/10.4_dbg/sql/sql_admin.cc:1036
 
#14 0x00005623bcc3c1ed in Sql_cmd_optimize_table::execute (this=<optimized out>, thd=0x14fbdec15070) at /test/10.4_dbg/sql/sql_admin.cc:1380
 
#15 0x00005623bcaf4a3d in mysql_execute_command (thd=thd@entry=0x14fbdec15070) at /test/10.4_dbg/sql/sql_parse.cc:6098
 
#16 0x00005623bcaf7090 in mysql_parse (thd=thd@entry=0x14fbdec15070, rawbuf=<optimized out>, length=<optimized out>, parser_state=parser_state@entry=0x14fc0635f460, is_com_multi=is_com_multi@entry=false, is_next_command=is_next_command@entry=false) at /test/10.4_dbg/sql/sql_parse.cc:7896
 
#17 0x00005623bcaf9920 in dispatch_command (command=command@entry=COM_QUERY, thd=thd@entry=0x14fbdec15070, packet=packet@entry=0x14fbdec57071 "", packet_length=packet_length@entry=16, is_com_multi=is_com_multi@entry=false, is_next_command=is_next_command@entry=false) at /test/10.4_dbg/sql/sql_parse.cc:1834
 
#18 0x00005623bcafd35b in do_command (thd=0x14fbdec15070) at /test/10.4_dbg/sql/sql_parse.cc:1352
 
#19 0x00005623bcc298b6 in do_handle_one_connection (connect=connect@entry=0x14fc03035790) at /test/10.4_dbg/sql/sql_connect.cc:1412
 
#20 0x00005623bcc299d6 in handle_one_connection (arg=0x14fc03035790) at /test/10.4_dbg/sql/sql_connect.cc:1316
 
#21 0x000014fc0555e6db in start_thread (arg=0x14fc06360700) at pthread_create.c:463
 
#22 0x000014fc046d8a3f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

10.4.15 eae968f62d285de97ed607c87bc131cd863d5d03 (Optimized)

 
Core was generated by `/test/MD110820-mariadb-10.4.15-linux-x86_64-opt/bin/mysqld --no-defaults --core'.
 
Program terminated with signal SIGSEGV, Segmentation fault.
 
#0  __pthread_kill (threadid=<optimized out>, signo=signo@entry=11)
 
    at ../sysdeps/unix/sysv/linux/pthread_kill.c:57
 
[Current thread is 1 (Thread 0x1511c881e700 (LWP 594389))]
 
(gdb) bt
 
#0  __pthread_kill (threadid=<optimized out>, signo=signo@entry=11) at ../sysdeps/unix/sysv/linux/pthread_kill.c:57
 
#1  0x0000562391ceba77 in my_write_core (sig=sig@entry=11) at /test/10.4_opt/mysys/stacktrace.c:482
 
#2  0x00005623916c362a in handle_fatal_signal (sig=11) at /test/10.4_opt/sql/signal_handler.cc:343
 
#3  <signal handler called>
 
#4  row_make_new_pathname (new_name=0x1511a14217c0 "test/#sql-ib22", table=0x1511a14842e8) at /test/10.4_opt/storage/innobase/row/row0merge.cc:4280
 
#5  row_merge_rename_tables_dict (old_table=old_table@entry=0x1511a14842e8, new_table=new_table@entry=0x1511a1484668, tmp_name=0x1511a14217c0 "test/#sql-ib22", trx=trx@entry=0x1511b30031e8) at /test/10.4_opt/storage/innobase/row/row0merge.cc:4335
 
#6  0x000056239188584f in commit_try_rebuild (table_name=<optimized out>, trx=<optimized out>, old_table=<optimized out>, altered_table=<optimized out>, ctx=0x1511a14405a8, ha_alter_info=<optimized out>) at /test/10.4_opt/storage/innobase/handler/handler0alter.cc:9944
 
#7  ha_innobase::commit_inplace_alter_table (this=0x1511a1464020, altered_table=<optimized out>, ha_alter_info=<optimized out>, commit=<optimized out>) at /test/10.4_opt/storage/innobase/handler/handler0alter.cc:11016
 
#8  0x0000562391549ad2 in mysql_inplace_alter_table (target_mdl_request=0x1511c8818ff0, alter_ctx=0x1511c881a100, ha_alter_info=0x1511c88180c0, altered_table=0x1511c8818160, table=0x1511a14cee08, table_list=0x1511a143f0e0, thd=0x1511a1412008) at /test/10.4_opt/sql/sql_table.cc:7797
 
#9  mysql_alter_table (thd=thd@entry=0x1511a1412008, new_db=<optimized out>, new_name=<optimized out>, create_info=create_info@entry=0x1511c881acb0, table_list=<optimized out>, table_list@entry=0x1511a143f0e0, alter_info=<optimized out>, order_num=0, order=0x0, ignore=false) at /test/10.4_opt/sql/sql_table.cc:10156
 
#10 0x000056239154aae8 in mysql_recreate_table (thd=thd@entry=0x1511a1412008, table_list=table_list@entry=0x1511a143f0e0, table_copy=table_copy@entry=false) at /test/10.4_opt/sql/sql_table.cc:11001
 
#11 0x00005623915a5c6d in admin_recreate_table (thd=thd@entry=0x1511a1412008, table_list=table_list@entry=0x1511a143f0e0) at /test/10.4_opt/sql/sql_admin.cc:59
 
#12 0x00005623915a7154 in mysql_admin_table (thd=thd@entry=0x1511a1412008, tables=tables@entry=0x1511a143f0e0, check_opt=check_opt@entry=0x1511a1417020, operator_name=operator_name@entry=0x562391f0a056 "optimize", lock_type=lock_type@entry=TL_WRITE, org_open_for_modify=org_open_for_modify@entry=true, repair_table_use_frm=false, extra_open_options=0, prepare_func=0x0, operator_func=<optimized out>, view_operator_func=0x0) at /test/10.4_opt/sql/sql_admin.cc:1036
 
#13 0x00005623915a8480 in Sql_cmd_optimize_table::execute (this=<optimized out>, thd=0x1511a1412008) at /test/10.4_opt/sql/sql_admin.cc:1380
 
#14 0x00005623914b2ca4 in mysql_execute_command (thd=thd@entry=0x1511a1412008) at /test/10.4_opt/sql/sql_parse.cc:6098
 
#15 0x00005623914ba22a in mysql_parse (thd=0x1511a1412008, rawbuf=<optimized out>, length=16, parser_state=0x1511c881d4d0, is_com_multi=<optimized out>, is_next_command=<optimized out>) at /test/10.4_opt/sql/sql_parse.cc:7896
 
#16 0x00005623914bc685 in dispatch_command (command=command@entry=COM_QUERY, thd=thd@entry=0x1511a1412008, packet=packet@entry=0x1511a1432009 "", packet_length=packet_length@entry=16, is_com_multi=is_com_multi@entry=false, is_next_command=is_next_command@entry=false) at /test/10.4_opt/sql/sql_parse.cc:1834
 
#17 0x00005623914bde04 in do_command (thd=0x1511a1412008) at /test/10.4_opt/sql/sql_parse.cc:1352
 
#18 0x000056239159adbe in do_handle_one_connection (connect=connect@entry=0x1511c5432748) at /test/10.4_opt/sql/sql_connect.cc:1412
 
#19 0x000056239159ae7d in handle_one_connection (arg=0x1511c5432748) at /test/10.4_opt/sql/sql_connect.cc:1316
 
#20 0x00001511c7a1c6db in start_thread (arg=0x1511c881e700) at pthread_create.c:463
 
#21 0x00001511c6b96a3f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

Bug confirmed present in:
MariaDB: 10.3.25 (dbg), 10.3.25 (opt), 10.4.15 (dbg), 10.4.15 (opt)

Bug confirmed not present in:
MariaDB: 10.1.47 (dbg), 10.1.47 (opt), 10.2.34 (dbg), 10.2.34 (opt), 10.5.6 (dbg), 10.5.6 (opt)
MySQL: 5.5.62 (dbg), 5.5.62 (opt), 5.6.47 (dbg), 5.6.47 (opt), 5.7.29 (dbg), 5.7.29 (opt), 8.0.19 (dbg), 8.0.19 (opt)

Comment by Marko Mäkelä [ 2020-09-22 ]

Even though I was not able to repeat a failure in 10.2, I applied the fix also there. The test innodb.alter_missing_tablespace covers the case where an AUTO_INCREMENT value of an existing table is changed without rebuilding the table. To fix this bug, a similar check had to be added to the table-rebuilding case. Other forms of native ALTER TABLE are deliberately allowed even though the tablespace is missing, to retain compatibility with MySQL 5.7.

Generated at Thu Feb 08 09:18:37 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.