Vladislav Vaintroub added a comment - 2018-10-11 16:38 - edited Thanks for the dump Note : debugger informs that there was a null pointer access in the below code SEL_ARG * Item_bool_func::get_mm_leaf(RANGE_OPT_PARAM *param, Field *field, KEY_PART *key_part, Item_func::Functype type, Item *value) { uint maybe_null=(uint) field->real_maybe_null(); <-- here, field is NULL pointer SEL_ARG *tree= 0; the key itself is being passed as second parameter key_part->field sel_arg= get_mm_leaf(param, key_part->field, key_part, type, value); inside Item_bool_func::get_mm_parts(RANGE_OPT_PARAM *param, Field *field, Item_func::Functype type, Item *value) Unfortunately, the dump does not contain heap, so it is not possible to look into the affected key_part

Elena Stepanova added a comment - 2018-11-11 23:47 sstamm , Are you sure you haven't over-edited the query? Obfuscating names and other sensible values is of course fine, but the structure should remain intact. In its current form, it just doesn't look plausible. Could it be that you lost brackets and/or some clauses? After checking it, please also provide the output of EXPLAIN EXTENDED <the correct query>; SHOW WARNINGS; and SHOW INDEX IN employee; SHOW INDEX IN absence; Thanks.

Sebastian Stamm added a comment - 2018-12-04 09:24 - edited switched to 10.3.11: EXPLAIN EXTENDED also crashes MariaDB. Table Non_unique Key_name Seq_in_index Column_name Collation Cardinality Sub_part Packed Null Index_type Comment Index_comment employee 0 PRIMARY 1 employeeId A 718       BTREE     employee 0 name 1 name A 718       BTREE     employee 0 exportName 1 exportName A 2     YES BTREE     employee 1 FK_employee_department 1 departmentId A 179       BTREE     employee 1 FK_employee_location 1 locationId A 22       BTREE     employee 1 FK_employee_unit 1 unitId A 20       BTREE     employee 1 FK_employee_skill 1 skillId A 79       BTREE     Table Non_unique Key_name Seq_in_index Column_name Collation Cardinality Sub_part Packed Null Index_type Comment Index_comment absence 0 PRIMARY 1 absenceId A 7715       BTREE     absence 1 FK_absence_employee 1 employeeId A 964       BTREE     absence 1 FK_absence_absencereason 1 absenceReasonId A 8       BTREE     Query is: SELECT absenceId FROM absence JOIN employee USING(employeeid) WHERE name = 'Test, Test' AND START <= '2018-12-04' AND END >= '2018-12-04' AND absenceId != 8711 Problem causes 'absenceId != 8711' When i remove this, the query works. Returns only the row with 8711, with the condition, no rows are expected. Error with 10.3.11; Server version: 10.3 . 11 -MariaDB key_buffer_size= 134217728 read_buffer_size= 67108864 max_used_connections= 13 max_threads= 65537 thread_count= 9 It is possible that mysqld could use up to key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 266027 K bytes of memory Hope that's ok; if not, decrease some variables in the equation.   Thread pointer: 0x53dc6c4038 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... mysqld.exe!Item_bool_func::get_mm_leaf()[opt_range.cc: 8047 ] mysqld.exe!Item_bool_func::get_mm_parts()[opt_range.cc: 7886 ] mysqld.exe!Item_bool_func2_with_rev::get_func_mm_tree()[item_cmpfunc.h: 461 ] mysqld.exe!Item_bool_func::get_full_func_mm_tree()[opt_range.cc: 7564 ] mysqld.exe!Item_bool_func::get_full_func_mm_tree_for_args()[item_cmpfunc.h: 201 ] mysqld.exe!Item_bool_func2_with_rev::get_mm_tree()[item_cmpfunc.h: 488 ] mysqld.exe!Item_cond_and::get_mm_tree()[opt_range.cc: 7619 ] mysqld.exe!calculate_cond_selectivity_for_table()[opt_range.cc: 3078 ] mysqld.exe!make_join_statistics()[sql_select.cc: 5000 ] mysqld.exe!JOIN::optimize_inner()[sql_select.cc: 1889 ] mysqld.exe!JOIN::optimize()[sql_select.cc: 1453 ] mysqld.exe!mysql_select()[sql_select.cc: 4225 ] mysqld.exe!mysql_explain_union()[sql_select.cc: 25868 ] mysqld.exe!execute_sqlcom_select()[sql_parse.cc: 6485 ] mysqld.exe!mysql_execute_command()[sql_parse.cc: 3767 ] mysqld.exe!mysql_parse()[sql_parse.cc: 8095 ] mysqld.exe!dispatch_command()[sql_parse.cc: 1852 ] mysqld.exe!do_command()[sql_parse.cc: 1394 ] mysqld.exe!threadpool_process_request()[threadpool_common.cc: 358 ] mysqld.exe!tp_callback()[threadpool_common.cc: 186 ] ntdll.dll!RtlFreeUnicodeString() ntdll.dll!RtlFreeUnicodeString() KERNEL32.DLL!BaseThreadInitThunk() ntdll.dll!RtlUserThreadStart()   Trying to get some variables. Some pointers may be invalid and cause the dump to abort. Query ( 0x70fee64030 ): EXPLAIN EXTENDED   SELECT absenceId   FROM absence   JOIN employee USING(employeeid)   WHERE name= 'Test, Test' AND START <= '2018-12-04' AND END >= '2018-12-04' AND absenceId != 8711 Connection ID (thread ID): 416 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=on,mrr_cost_based=on,mrr_sort_keys=on,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

Alice Sherepa added a comment - 2019-01-07 13:14 - edited Do this tables contain confidential information? Would you be able to upload datadump of them to ftp.askmonty.org/private?

Elena Stepanova added a comment - 2019-02-14 14:12 alice , it looks much like MDEV-18300 , except that I don't see any BLOBs here; but possibly with other configuration options mentioned here the test case from MDEV-18300 can be tweaked to avoid blobs?