|
MTR test case looks like this:
--source include/have_innodb.inc
|
|
--let MYSQLD_DATADIR= `select @@datadir`
|
|
SET SESSION sql_mode = 'NO_ENGINE_SUBSTITUTION' ;
|
SET DEFAULT_STORAGE_ENGINE= 'InnoDB' ;
|
USE test ;
|
CREATE TABLE `t1` (
|
`col_int` int,
|
`col_varchar_255` varchar(255),
|
`col_text` text) ENGINE=innodb ;
|
INSERT /*! IGNORE */ INTO t1 VALUES
|
(1028915200, 'c', 'image') ,
|
(NULL, 'j', 'historic') ,
|
(6, 'loyal', 'retreat') ,
|
(NULL, 'depth', 'male') ,
|
(1, 'squad', 'too') ,
|
(-1846607872, 'roof', 'head') ,
|
(NULL, 'o', 'turkey') ,
|
(NULL, 'w', 'deal') ,
|
(3, 'f', 'cousin') ,
|
(-996605952, 'i', 'online') ;
|
FLUSH TABLES t1 FOR EXPORT ;
|
UNLOCK TABLES ;
|
|
--copy_file $MYSQLD_DATADIR/test/t1.ibd $MYSQLD_DATADIR/test/tmp.ibd
|
|
CREATE TABLE imp_t1 LIKE t1 ;
|
ALTER TABLE imp_t1 DISCARD TABLESPACE ;
|
ALTER TABLE imp_t1 MODIFY COLUMN col_varchar_255 VARCHAR(255) FIRST ;
|
ALTER TABLE imp_t1 DISCARD TABLESPACE ;
|
--copy_file $MYSQLD_DATADIR/test/tmp.ibd $MYSQLD_DATADIR/test/imp_t1.ibd
|
ALTER TABLE imp_t1 IMPORT TABLESPACE ;
|
ALTER TABLE imp_t1 MODIFY COLUMN col_int INT FIRST ;
|
ALTER TABLE imp_t1 MODIFY COLUMN col_varchar_255 VARCHAR(255) FIRST ;
|
|
|
|
kevg, thank you. Nitpick: the first copy_file should be between FLUSH TABLES…FOR EXPORT and UNLOCK TABLES.
The test case appears to simulate a user error: importing a table where the columns are in the wrong order, while not using a .cfg file (so that this metadata mismatch could be caught).
|
|
with Eugene's test:
|
10.6 892c426371b4be558d32fdeb
|
|
2022-06-09 17:18:49 4 [ERROR] InnoDB: Page old data size 418 new data size 408, page old max ins size 15828 new max ins size 15838
|
mariadbd: /10.6/src/storage/innobase/include/rem0rec.h:274: rec_comp_status_t rec_get_status(const rec_t*): Assertion `bits <= REC_STATUS_INSTANT' failed.
|
220609 17:18:49 [ERROR] mysqld got signal 6 ;
|
|
Server version: 10.6.9-MariaDB-debug-log
|
|
??:0(__assert_fail)[0x7f1a6c49bfd6]
|
include/rem0rec.h:275(rec_get_status(unsigned char const*))[0x558e4b7e7e58]
|
rem/rem0rec.cc:851(rec_get_offsets_func(unsigned char const*, dict_index_t const*, unsigned short*, unsigned long, unsigned long, char const*, unsigned int, mem_block_info_t**))[0x558e4ba1a7e1]
|
handler/handler0alter.cc:5988(innobase_instant_try(Alter_inplace_info const*, ha_innobase_inplace_ctx*, TABLE const*, TABLE const*, trx_t*))[0x558e4b7b1686]
|
handler/handler0alter.cc:10448(commit_try_norebuild(Alter_inplace_info*, ha_innobase_inplace_ctx*, TABLE*, TABLE const*, trx_t*, char const*))[0x558e4b8055fb]
|
handler/handler0alter.cc:11196(ha_innobase::commit_inplace_alter_table(TABLE*, Alter_inplace_info*, bool))[0x558e4b7de091]
|
sql/handler.cc:5220(handler::ha_commit_inplace_alter_table(TABLE*, Alter_inplace_info*, bool))[0x558e4ac2c550]
|
sql/sql_table.cc:7471(mysql_inplace_alter_table(THD*, TABLE_LIST*, TABLE*, TABLE*, Alter_inplace_info*, MDL_request*, st_ddl_log_state*, TRIGGER_RENAME_PARAM*, Alter_table_ctx*))[0x558e4a632a6a]
|
sql/sql_table.cc:10278(mysql_alter_table(THD*, st_mysql_const_lex_string const*, st_mysql_const_lex_string const*, HA_CREATE_INFO*, TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool, bool))[0x558e4a647e7c]
|
sql/sql_alter.cc:542(Sql_cmd_alter_table::execute(THD*))[0x558e4a8098c8]
|
sql/sql_parse.cc:5996(mysql_execute_command(THD*, bool))[0x558e4a3a176c]
|
sql/sql_parse.cc:8029(mysql_parse(THD*, char*, unsigned int, Parser_state*))[0x558e4a3aee48]
|
sql/sql_parse.cc:1898(dispatch_command(enum_server_command, THD*, char*, unsigned int, bool))[0x558e4a38509c]
|
sql/sql_parse.cc:1409(do_command(THD*, bool))[0x558e4a381e11]
|
sql/sql_connect.cc:1418(do_handle_one_connection(CONNECT*, bool))[0x558e4a7ec100]
|
sql/sql_connect.cc:1314(handle_one_connection)[0x558e4a7eb98c]
|
perfschema/pfs.cc:2203(pfs_spawn_thread)[0x558e4b51ce36]
|
nptl/pthread_create.c:478(start_thread)[0x7f1a6c9b6609]
|
addr2line: DWARF error: section .debug_info is larger than its filesize! (0x93ef57 vs 0x530ea0)
|
??:0(clone)[0x7f1a6c587133]
|
|
Query (0x62b0000c42a8): ALTER TABLE imp_t1 MODIFY COLUMN col_varchar_255 VARCHAR(255) FIRST
|
on non-debug:
2022-06-09 17:24:10 4 [ERROR] [FATAL] InnoDB: Page old data size 418 new data size 408, page old max ins size 15828 new max ins size 15838
|
220609 17:24:10 [ERROR] mysqld got signal 6 ;
|
|
Server version: 10.6.7-MariaDB
|
|
??:0(gsignal)[0x7fcd5581d00b]
|
??:0(abort)[0x7fcd557fc859]
|
ut/ut0vec.cc:38(ib_vector_create(ib_alloc_t*, unsigned long, unsigned long))[0x5621dfc43f70]
|
btr/btr0btr.cc:1266(btr_page_reorganize_low(page_cur_t*, dict_index_t*, mtr_t*) [clone .isra.88])[0x5621dfc4a46c]
|
btr/btr0btr.cc:1565(btr_page_reorganize(page_cur_t*, dict_index_t*, mtr_t*))[0x5621dfc4a877]
|
btr/btr0cur.cc:5126(btr_cur_pessimistic_update(unsigned long, btr_cur_t*, unsigned short**, mem_block_info_t**, mem_block_info_t*, big_rec_t**, upd_t*, unsigned long, que_thr_t*, unsigned long, mtr_t*))[0x5621dfc6bf24]
|
handler/handler0alter.cc:5904(commit_try_norebuild(Alter_inplace_info*, ha_innobase_inplace_ctx*, TABLE*, TABLE const*, trx_t*, char const*))[0x5621dfb23030]
|
handler/handler0alter.cc:11067(ha_innobase::commit_inplace_alter_table(TABLE*, Alter_inplace_info*, bool))[0x5621dfb1b0be]
|
sql/sql_table.cc:7469(mysql_alter_table(THD*, st_mysql_const_lex_string const*, st_mysql_const_lex_string const*, HA_CREATE_INFO*, TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool, bool))[0x5621df662058]
|
sql/sql_alter.cc:548(Sql_cmd_alter_table::execute(THD*))[0x5621df6c7b6e]
|
sql/sql_parse.cc:5996(mysql_execute_command(THD*, bool))[0x5621df5c4d75]
|
sql/sql_parse.cc:8030(mysql_parse(THD*, char*, unsigned int, Parser_state*))[0x5621df5c959b]
|
sql/sql_parse.cc:1955(dispatch_command(enum_server_command, THD*, char*, unsigned int, bool))[0x5621df5cb7e7]
|
sql/sql_parse.cc:1406(do_command(THD*, bool))[0x5621df5cce83]
|
sql/sql_connect.cc:1418(do_handle_one_connection(CONNECT*, bool))[0x5621df6c2be7]
|
sql/sql_connect.cc:1318(handle_one_connection)[0x5621df6c2e84]
|
perfschema/pfs.cc:2204(pfs_spawn_thread)[0x5621dfa560cc]
|
nptl/pthread_create.c:478(start_thread)[0x7fcd55d0d609]
|
|
|
Query (0x7fccf8012210): ALTER TABLE imp_t1 MODIFY COLUMN col_varchar_255 VARCHAR(255) FIRST
|
|
|
alice, this looks like a duplicate of MDEV-18519. Can you please double-check the revision that you tested? 892c426371b4be558d32fdeb should not crash there, but issue an error message instead.
|
|
repeated on 10.6 51ca5d517ef67f97e3c38d57174f8
|
|
alice, 10.6 51ca5d517ef67f97e3c38d57174f8 is after MDEV-13542, which should have replaced the crash with an error report. Can you please post the details for the crash that you reproduced?
|
|
marko, I ran mtr test, that was provided in the first comment by Eugene Kosov. Repeated now also on 10.6 b9eb63618edcad6206, built with -DWITH_ASAN=ON
2022-07-26 15:18:00 4 [ERROR] InnoDB: Page old data size 418 new data size 408, page old max ins size 15828 new max ins size 15838
|
mariadbd: /10.6/src/storage/innobase/include/rem0rec.h:274: rec_comp_status_t rec_get_status(const rec_t*): Assertion `bits <= REC_STATUS_INSTANT' failed.
|
220726 15:18:00 [ERROR] mysqld got signal 6 ;
|
This could be because you hit a bug. It is also possible that this binary
|
or one of the libraries it was linked against is corrupt, improperly built,
|
or misconfigured. This error can also be caused by malfunctioning hardware.
|
|
To report this bug, see https://mariadb.com/kb/en/reporting-bugs
|
|
We will try our best to scrape up some info that will hopefully help
|
diagnose the problem, but since we have already crashed,
|
something is definitely wrong and this may fail.
|
|
Server version: 10.6.9-MariaDB-debug-log
|
key_buffer_size=1048576
|
read_buffer_size=131072
|
max_used_connections=1
|
max_threads=153
|
thread_count=1
|
It is possible that mysqld could use up to
|
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 63859 K bytes of memory
|
Hope that's ok; if not, decrease some variables in the equation.
|
|
Thread pointer: 0x62b0000bd288
|
Attempting backtrace. You can use the following information to find out
|
where mysqld died. If you see no messages after this, something went
|
terribly wrong...
|
stack_bottom = 0x7fec4f852c50 thread_stack 0x100000
|
sanitizer_common/sanitizer_common_interceptors.inc:4022(__interceptor_backtrace.part.0)[0x7fec5ec55d40]
|
mysys/stacktrace.c:212(my_print_stacktrace)[0x56111fdf281f]
|
sql/signal_handler.cc:236(handle_fatal_signal)[0x56111e972496]
|
sigaction.c:0(__restore_rt)[0x7fec5e6e7420]
|
??:0(gsignal)[0x7fec5e1d000b]
|
??:0(abort)[0x7fec5e1af859]
|
/lib/x86_64-linux-gnu/libc.so.6(+0x22729)[0x7fec5e1af729]
|
??:0(__assert_fail)[0x7fec5e1c0fd6]
|
include/rem0rec.h:275(rec_get_status(unsigned char const*))[0x56111f4a224e]
|
rem/rem0rec.cc:851(rec_get_offsets_func(unsigned char const*, dict_index_t const*, unsigned short*, unsigned long, unsigned long, char const*, unsigned int, mem_block_info_t**))[0x56111f6d9b19]
|
handler/handler0alter.cc:6115(innobase_instant_try(Alter_inplace_info const*, ha_innobase_inplace_ctx*, TABLE const*, TABLE const*, trx_t*))[0x56111f46bd3a]
|
handler/handler0alter.cc:10548(commit_try_norebuild(Alter_inplace_info*, ha_innobase_inplace_ctx*, TABLE*, TABLE const*, trx_t*, char const*))[0x56111f4c1027]
|
handler/handler0alter.cc:11297(ha_innobase::commit_inplace_alter_table(TABLE*, Alter_inplace_info*, bool))[0x56111f497edf]
|
sql/handler.cc:5220(handler::ha_commit_inplace_alter_table(TABLE*, Alter_inplace_info*, bool))[0x56111e9a0478]
|
sql/sql_table.cc:7467(mysql_inplace_alter_table(THD*, TABLE_LIST*, TABLE*, TABLE*, Alter_inplace_info*, MDL_request*, st_ddl_log_state*, TRIGGER_RENAME_PARAM*, Alter_table_ctx*))[0x56111e3a5e92]
|
sql/sql_table.cc:10274(mysql_alter_table(THD*, st_mysql_const_lex_string const*, st_mysql_const_lex_string const*, HA_CREATE_INFO*, TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool, bool))[0x56111e3bb2a4]
|
sql/sql_alter.cc:542(Sql_cmd_alter_table::execute(THD*))[0x56111e57ccc2]
|
sql/sql_parse.cc:5996(mysql_execute_command(THD*, bool))[0x56111e114b0a]
|
sql/sql_parse.cc:8029(mysql_parse(THD*, char*, unsigned int, Parser_state*))[0x56111e1221e6]
|
sql/sql_parse.cc:1898(dispatch_command(enum_server_command, THD*, char*, unsigned int, bool))[0x56111e0f843a]
|
sql/sql_parse.cc:1409(do_command(THD*, bool))[0x56111e0f51af]
|
sql/sql_connect.cc:1418(do_handle_one_connection(CONNECT*, bool))[0x56111e55f4fa]
|
sql/sql_connect.cc:1314(handle_one_connection)[0x56111e55ed86]
|
perfschema/pfs.cc:2203(pfs_spawn_thread)[0x56111f1d787e]
|
nptl/pthread_create.c:478(start_thread)[0x7fec5e6db609]
|
??:0(clone)[0x7fec5e2ac133]
|
|
Query (0x62b0000c42a8): ALTER TABLE imp_t1 MODIFY COLUMN col_varchar_255 VARCHAR(255) FIRST
|
|
|
Another occurrence of this issue:
The first crash happened during an UPDATE DML: `update t set a=2,b=now() where id=38479845`
Error log of the first crash:
2022-07-26 11:14:25 22804036 [ERROR] [FATAL] InnoDB: Page old data size 15162 new data size 6230, page old max ins size 867 new max ins size 9799
|
220726 11:14:25 [ERROR] mysqld got signal 6 ;
|
This could be because you hit a bug. It is also possible that this binary
|
or one of the libraries it was linked against is corrupt, improperly built,
|
or misconfigured. This error can also be caused by malfunctioning hardware.
|
|
To report this bug, see https://mariadb.com/kb/en/reporting-bugs
|
|
We will try our best to scrape up some info that will hopefully help
|
diagnose the problem, but since we have already crashed,
|
something is definitely wrong and this may fail.
|
|
Server version: 10.5.13-MariaDB-log
|
key_buffer_size=16777216
|
read_buffer_size=262144
|
max_used_connections=460
|
max_threads=2626
|
thread_count=358
|
It is possible that mysqld could use up to
|
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 6091296 K bytes of memory
|
Hope that's ok; if not, decrease some variables in the equation.
|
|
Thread pointer: 0x400c87333758
|
Attempting backtrace. You can use the following information to find out
|
where mysqld died. If you see no messages after this, something went
|
terribly wrong...
|
stack_bottom = 0x400c3f0c6060 thread_stack 0x40000
|
/bin/mysqld(my_print_stacktrace+0x2c)[0xaaaabaf2bfb8]
|
|
Trying to get some variables.
|
Some pointers may be invalid and cause the dump to abort.
|
Query (0x400c8735b2f0): update t set a=2,b=now() where id=38479845
|
|
Connection ID (thread ID): 22804036
|
Status: NOT_KILLED
|
|
Optimizer switch: index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=on,table_elimination=on,extended_keys=on,exists_to_in=on,orderby_uses_equalities=on,condition_pushdown_for_derived=on,split_materialized=on,condition_pushdown_for_subquery=on,rowid_filter=on,condition_pushdown_from_having=on,not_null_range_scan=off
|
|
The manual page at https://mariadb.com/kb/en/how-to-produce-a-full-stack-trace-for-mysqld/ contains
|
information that should help you find out what is causing the crash.
|
Writing a core file...
|
Working directory at /db
|
Resource Limits:
|
Limit Soft Limit Hard Limit Units
|
Max cpu time unlimited unlimited seconds
|
Max file size unlimited unlimited bytes
|
Max data size unlimited unlimited bytes
|
Max stack size unlimited unlimited bytes
|
Max core file size 0 unlimited bytes
|
Max resident set unlimited unlimited bytes
|
Max processes unlimited unlimited processes
|
Max open files 65535 65535 files
|
Max locked memory 65536 65536 bytes
|
Max address space unlimited unlimited bytes
|
Max file locks unlimited unlimited locks
|
Max pending signals 252382 252382 signals
|
Max msgqueue size 819200 819200 bytes
|
Max nice priority 0 0
|
Max realtime priority 0 0
|
Max realtime timeout unlimited unlimited us
|
Core pattern: /tmp/core-%e-%p
|
Then the server stuck in crash loop and couldn't start:
Stack trace of the crashing thread now, caller functions are different with the one in Jira description, but both in function `btr_page_reorganize`:
|
Thread 40 (Thread 0x400c5d7fe910 (LWP 25198)):
|
#0 0x000040001e91d4f0 in raise () from /lib64/libc.so.6
|
#1 0x000040001e91ed34 in abort () from /lib64/libc.so.6
|
#2 0x0000aaaaddb0837c in ib::fatal::~fatal (this=<optimized out>, __in_chrg=<optimized out>) at /MariaDB/storage/innobase/ut/ut0ut.cc:576
|
#3 0x0000aaaaddb0ed04 in btr_page_reorganize_low (index=index@entry=0x400c5d808800, mtr=mtr@entry=0x400c5d7faf08, cursor=<optimized out>, cursor=<optimized out>) at /MariaDB/storage/innobase/btr/btr0btr.cc:1411
|
#4 0x0000aaaaddb0f044 in btr_page_reorganize (cursor=cursor@entry=0x400c5d7f9fa0, index=index@entry=0x400c5d808800, mtr=mtr@entry=0x400c5d7faf08) at /MariaDB/storage/innobase/btr/btr0btr.cc:1629
|
#5 0x0000aaaadda07c7c in ibuf_insert_to_index_page_low (entry=0x400c5d808610, entry@entry=0x0, block=block@entry=0x4001337cc300, index=index@entry=0x400c5d808800, offsets=0x400c5d7f9f28, offsets@entry=0x400c5d7f9fe8, heap=<optimized out>, mtr=mtr@entry=0x400c5d7faf08,
|
page_cur=page_cur@entry=0x400c5d7f9fa0) at /MariaDB/storage/innobase/ibuf/ibuf0ibuf.cc:3680
|
#6 0x0000aaaadda0ca0c in ibuf_insert_to_index_page (mtr=0x400c5d7fafc8, index=0x400c5d808800, block=0x4001337cc300, entry=0x0) at /MariaDB/storage/innobase/ibuf/ibuf0ibuf.cc:3881
|
#7 ibuf_merge_or_delete_for_page (block=block@entry=0x4001337cc300, page_id=..., zip_size=0, zip_size@entry=187650851803136) at /MariaDB/storage/innobase/ibuf/ibuf0ibuf.cc:4425
|
#8 0x0000aaaaddb3d66c in buf_page_get_low (page_id=..., zip_size=187650851803136, zip_size@entry=0, rw_latch=70372876156928, rw_latch@entry=2, guess=<optimized out>, mode=70372828062976, file=0x15 <error: Cannot access memory at address 0x15>, line=43690,
|
mtr=0x400c5d7fb6e0, err=err@entry=0x400c5d7fb8b4, allow_ibuf_merge=allow_ibuf_merge@entry=true) at /MariaDB/storage/innobase/buf/buf0buf.cc:3382
|
#9 0x0000aaaaddb3d6d4 in buf_page_get_gen (page_id=page_id@entry=..., zip_size=0, zip_size@entry=187650835869280, rw_latch=rw_latch@entry=2, guess=guess@entry=0x0, mode=mode@entry=10, file=file@entry=0xaaaadd69e9e8 <extent_split_impl+96> "\373\003", line=line@entry=360,
|
mtr=mtr@entry=0x9, err=0x400c5d7fb8b4, err@entry=0x400c5d7fba80, allow_ibuf_merge=<optimized out>) at /MariaDB/storage/innobase/buf/buf0buf.cc:3464
|
#10 0x0000aaaaddb204f4 in btr_cur_search_to_nth_level_func (index=index@entry=0x40001f368000, level=70421852437120, level@entry=0, tuple=0x3a0efff, tuple@entry=0x400c5d838080, mode=PAGE_CUR_UNSUPP, mode@entry=PAGE_CUR_LE, latch_mode=latch_mode@entry=2,
|
cursor=0x400c5d7fba80, cursor@entry=0x400c5d7fc760, ahi_latch=0x4001337a8f08, ahi_latch@entry=0x0, file=0xaaaadd69e9e8 <extent_split_impl+96> "\373\003",
|
file@entry=0xaaaaddfe9dc0 "/MariaDB/storage/innobase/row/row0row.cc", line=line@entry=1303, mtr=0x400c5d7fc9c8, autoinc=autoinc@entry=0)
|
at /MariaDB/storage/innobase/btr/btr0cur.cc:1616
|
#11 0x0000aaaaddaa21c8 in btr_pcur_open_low (index=index@entry=0x40001f368000, tuple=tuple@entry=0x400c5d838080, latch_mode=latch_mode@entry=2, cursor=cursor@entry=0x400c5d7fc760, line=line@entry=1303, mtr=mtr@entry=0x400c5d7fc9c8, autoinc=0,
|
file=0xaaaaddfe9dc0 "/MariaDB/storage/innobase/row/row0row.cc", mode=PAGE_CUR_LE, level=0) at /MariaDB/storage/innobase/include/btr0pcur.ic:441
|
#12 0x0000aaaaddaa2318 in row_search_index_entry (index=index@entry=0x40001f368000, entry=entry@entry=0x400c5d838080, mode=2, pcur=pcur@entry=0x400c5d7fc760, mtr=mtr@entry=0x400c5d7fc9c8)
|
at /MariaDB/storage/innobase/row/row0row.cc:1303
|
#13 0x0000aaaaddbfba0c in row_undo_mod_del_mark_or_remove_sec_low (node=node@entry=0x400c5d814080, thr=thr@entry=0x400c5d8083b0, index=index@entry=0x40001f368000, entry=entry@entry=0x400c5d838080, mode=<optimized out>, mode@entry=2)
|
at /MariaDB/storage/innobase/row/row0umod.cc:565
|
#14 0x0000aaaaddbfd84c in row_undo_mod_del_mark_or_remove_sec (entry=0x400c5d838080, index=<optimized out>, thr=0x400c5d8083b0, node=0x400c5d814080) at /MariaDB/storage/innobase/row/row0umod.cc:670
|
#15 row_undo_mod_upd_exist_sec (thr=0x400c5d8083b0, node=0x400c5d814080) at /MariaDB/storage/innobase/row/row0umod.cc:1148
|
#16 row_undo_mod (node=node@entry=0x400c5d814080, thr=thr@entry=0x400c5d8083b0) at /MariaDB/storage/innobase/row/row0umod.cc:1363
|
#17 0x0000aaaaddaaafec in row_undo (thr=0x400c5d8083b0, node=0x400c5d814080) at /MariaDB/storage/innobase/row/row0undo.cc:424
|
#18 row_undo_step (thr=thr@entry=0x400c5d8083b0) at /MariaDB/storage/innobase/row/row0undo.cc:476
|
#19 0x0000aaaadda5d4c8 in que_thr_step (thr=0x400c5d8083b0) at /MariaDB/storage/innobase/que/que0que.cc:944
|
#20 que_run_threads_low (thr=0x400c5d8083b0) at /MariaDB/storage/innobase/que/que0que.cc:1008
|
#21 que_run_threads (thr=0x400c5d8083b0) at /MariaDB/storage/innobase/que/que0que.cc:1048
|
#22 0x0000aaaaddae43b0 in trx_rollback_active (trx=trx@entry=0x40002b386e50) at /MariaDB/storage/innobase/trx/trx0roll.cc:628
|
#23 0x0000aaaaddae497c in trx_rollback_recovered (all=all@entry=true) at /MariaDB/storage/innobase/trx/trx0roll.cc:783
|
#24 0x0000aaaaddae5010 in trx_rollback_all_recovered () at /MariaDB/storage/innobase/trx/trx0roll.cc:837
|
#25 0x000040001e8bd22c in start_thread () from /lib64/libpthread.so.0
|
#26 0x000040001e9bde5c in thread_start () from /lib64/libc.so.6
|
|
|
|
|
Thank you, alice, I can repeat this with that test.
|
10.6 63478e72de6940abae6bb2d3b35c9b1ffa7180d9
|
mysqltest: At line 35: query 'ALTER TABLE imp_t1 MODIFY COLUMN col_varchar_255 VARCHAR(255) FIRST ' failed: <Unknown> (2013): Lost connection to server during query
|
…
|
#3 0x00007ffb80658242 in __GI___assert_fail (assertion=0x562e98781360 "bits <= REC_STATUS_INSTANT", file=0x562e987813a0 "/mariadb/10.6/storage/innobase/include/rem0rec.h", line=253,
|
function=0x562e98781400 "rec_comp_status_t rec_get_status(const rec_t*)") at assert.c:101
|
#4 0x0000562e97a71655 in rec_get_status (rec=rec@entry=0x7ffb7dd781f3 "") at /mariadb/10.6/storage/innobase/include/rem0rec.h:253
|
#5 0x0000562e97c28eec in rec_get_offsets_func (rec=0x7ffb7dd781f3 "", index=index@entry=0x61600097b108, offsets=offsets@entry=0x615000034688, n_core=n_core@entry=6,
|
n_fields=n_fields@entry=18446744073709551615, file=file@entry=0x562e9877a2e0 "/mariadb/10.6/storage/innobase/handler/handler0alter.cc", line=6123, heap=0x7ffb766c9eb0)
|
at /mariadb/10.6/storage/innobase/rem/rem0rec.cc:851
|
#6 0x0000562e97a98ff2 in innobase_instant_try (ha_alter_info=ha_alter_info@entry=0x7ffb766cbc50, ctx=ctx@entry=0x62b0000b7ca8, altered_table=altered_table@entry=0x7ffb766cc7d0,
|
table=table@entry=0x6190000d7a98, trx=trx@entry=0x7ffb7e0fce40) at /mariadb/10.6/storage/innobase/handler/handler0alter.cc:6123
|
The rec address passed to rec_get_status() would have been valid (the hidden metadata record for instant ADD/DROP) before btr_page_reorganize_low() rebuilt the page during btr_cur_pessimistic_update(). After page_copy_rec_list_end_no_locks(), the metadata record would be at byte offset 0x71 from the start of the page, instead of the old offset 0x1f3. The cursor position was not adjusted due to the detected and reported size mismatch.
If we ignore the root cause of this bug (the size mismatch) for a while, in 10.6 there is a problem that the error handling after MDEV-13542 could be improved. With a couple of minor things fixed, it would still crash when attempting to roll back the ALTER TABLE transaction:
|
10.6 34cdc0062882ddfaf48bad0c079266d8c00c486c
|
2022-08-01 16:26:25 4 [ERROR] InnoDB: Page old data size 418 new data size 408, page old max ins size 15828 new max ins size 15838
|
2022-08-01 16:26:25 0x7f2160d60640 InnoDB: Assertion failure in file /mariadb/10.6/storage/innobase/btr/btr0cur.cc line 4520
|
InnoDB: Failing assertion: !rec_offs_any_null_extern(rec, *offsets) || thr_get_trx(thr) == trx_roll_crash_recv_trx
|
wenhug, the stack trace in your crash must be something else. It looks like ROLLBACK of some DML. Please note that disabling the change buffer will not instantly fix all dormant corruption due to the change buffer; it will only prevent the future use of the change buffer. See MDEV-27734 and related bugs.
This scenario of this bug is strictly related to MDEV-15562.
|
|
Here is a simpler test case, better illustrating a design problem that MariaDB inherited from MySQL 4.1:
--source include/have_innodb.inc
|
--let MYSQLD_DATADIR= `select @@datadir`
|
CREATE TABLE t1 (a INT, b varchar(255), c text) ENGINE=innodb ;
|
INSERT INTO t1 VALUES(NULL,'','');
|
FLUSH TABLES t1 FOR EXPORT;
|
--copy_file $MYSQLD_DATADIR/test/t1.ibd $MYSQLD_DATADIR/test/tmp.ibd
|
--copy_file $MYSQLD_DATADIR/test/t1.cfg $MYSQLD_DATADIR/test/tmp.cfg
|
UNLOCK TABLES;
|
CREATE TABLE imp_t1 LIKE t1 ;
|
ALTER TABLE imp_t1 DISCARD TABLESPACE ;
|
ALTER TABLE imp_t1 MODIFY COLUMN b VARCHAR(255) FIRST;
|
--move_file $MYSQLD_DATADIR/test/tmp.ibd $MYSQLD_DATADIR/test/imp_t1.ibd
|
#--move_file $MYSQLD_DATADIR/test/tmp.cfg $MYSQLD_DATADIR/test/imp_t1.cfg
|
ALTER TABLE imp_t1 IMPORT TABLESPACE ;
|
ALTER TABLE imp_t1 MODIFY COLUMN a INT FIRST;
|
ALTER TABLE imp_t1 MODIFY COLUMN b VARCHAR(255) FIRST;
|
If the INSERT or the MODIFY COLUMN a is removed, the test will not crash. Because we are not providing a .cfg file like we should, the metadata mismatch that was deliberately caused between DISCARD and {IMPORT}} cannot be detected at IMPORT time. If the last move_file is enabled, the test will fail like this:
|
10.6 84d26f98c7f2007d9414c33dcbfdddd3a24c9a60
|
mysqltest: At line 14: query 'ALTER TABLE imp_t1 IMPORT TABLESPACE ' failed: ER_TABLE_SCHEMA_MISMATCH (1808): Schema mismatch (Column b ordinal value mismatch, it's at 0 in the table and 1 in the tablespace meta-data file)
|
We really need MDEV-11658 and have to get rid of the broken-by-design DISCARD TABLESPACE workflow.
|
|
Possibly related to MDEV-18322
|
|
MDEV-18322 is something different, not related to importing a tablespace that is inconsistent with the table definition.
|
ML_import5.test