Note: The problems in the test case start on the first table already. SHOW TABLE STATUS returns Warning 1030 Got error 174 "Fatal error during initialization of handler" from storage engine Aria. This part is described in MDEV-17968 (unrelated to versioning). However, it is disturbing that even after being dropped, this corrupt table keeps affecting further actions.
--source include/have_partition.inc
|
|
CREATE TABLE t1 (a INT PRIMARY KEY) ENGINE=Aria PAGE_CHECKSUM=0 WITH SYSTEM VERSIONING PARTITION BY SYSTEM_TIME (PARTITION p1 HISTORY, PARTITION pn CURRENT);
|
ALTER TABLE t1 ADD PARTITION (PARTITION p2 HISTORY);
|
SHOW TABLE STATUS;
|
DROP TABLE t1;
|
|
CREATE TABLE t1 (b INT) ENGINE=Aria WITH SYSTEM VERSIONING ROW_FORMAT=DYNAMIC PARTITION BY SYSTEM_TIME (PARTITION p1 HISTORY, PARTITION pn CURRENT);
|
INSERT INTO t1 VALUES (1);
|
REPLACE INTO t1 VALUES (1);
|
|
# Cleanup
|
DROP TABLE t1;
|
On two of machines where I tried, it reliably crashes upon REPLACE on all of debug, non-debug and ASAN builds as follows:
|
10.4 af912664
|
#3 <signal handler called>
|
#4 key_copy (to_key=to_key@entry=0x7fd4d434a730 "", from_record=0x7fd4b8082430 "\376\001", key_info=0x0, key_length=key_length@entry=0, with_zerofill=with_zerofill@entry=false) at /data/src/10.4/sql/key.cc:122
|
#5 0x000055f1fab83773 in write_record (thd=thd@entry=0x7fd4b8000c08, table=table@entry=0x7fd4b811e798, info=info@entry=0x7fd4d434a8b0) at /data/src/10.4/sql/sql_insert.cc:1804
|
#6 0x000055f1fab89896 in mysql_insert (thd=thd@entry=0x7fd4b8000c08, table_list=<optimized out>, fields=..., values_list=..., update_fields=..., update_values=..., duplic=<optimized out>, ignore=<optimized out>) at /data/src/10.4/sql/sql_insert.cc:1078
|
#7 0x000055f1fabb535a in mysql_execute_command (thd=0x7fd4b8000c08) at /data/src/10.4/sql/sql_parse.cc:4528
|
#8 0x000055f1fabbbfd8 in mysql_parse (thd=thd@entry=0x7fd4b8000c08, rawbuf=<optimized out>, length=26, parser_state=parser_state@entry=0x7fd4d434d190, is_com_multi=is_com_multi@entry=false, is_next_command=is_next_command@entry=false) at /data/src/10.4/sql/sql_parse.cc:7900
|
#9 0x000055f1fabbe2af in dispatch_command (command=COM_QUERY, thd=0x7fd4b8000c08, packet=<optimized out>, packet_length=<optimized out>, is_com_multi=<optimized out>, is_next_command=<optimized out>) at /data/src/10.4/sql/sql_class.h:1168
|
#10 0x000055f1fabbf9df in do_command (thd=0x7fd4b8000c08) at /data/src/10.4/sql/sql_parse.cc:1359
|
#11 0x000055f1fac938f4 in do_handle_one_connection (connect=connect@entry=0x55f1fd5f9218) at /data/src/10.4/sql/sql_connect.cc:1412
|
#12 0x000055f1fac939e4 in handle_one_connection (arg=arg@entry=0x55f1fd5f9218) at /data/src/10.4/sql/sql_connect.cc:1316
|
#13 0x000055f1fb1ed2bf in pfs_spawn_thread (arg=0x55f1fd608d38) at /data/src/10.4/storage/perfschema/pfs.cc:1869
|
#14 0x00007fd4d9c48fa3 in start_thread (arg=<optimized out>) at pthread_create.c:486
|
#15 0x00007fd4d94974cf in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
|
On the third machine, it fails earlier, upon the 2nd CREATE TABLE, with an error (also on all of debug, non-debug and ASAN builds):
|
10.4 af912664
|
mysqltest: At line 8: query 'CREATE TABLE t1 (b INT) ENGINE=Aria WITH SYSTEM VERSIONING ROW_FORMAT=DYNAMIC PARTITION BY SYSTEM_TIME (PARTITION p1 HISTORY, PARTITION pn CURRENT)' failed: 156: Aria table 't1#P#pn' is in use (most likely by a MERGE table). Try FLUSH TABLES.
|
(FLUSH TABLES doesn't help)
Reproducible on 10.4, 10.5. Not reproducible on 10.3.
Bug
Major
