mirror of
https://github.com/MariaDB/server.git
synced 2025-01-31 02:51:44 +01:00
a6f258e47f
Backported from 10.520e9e804c1and5948d7602e. sel_restore_position_for_mysql() moves forward persistent cursor position after btr_pcur_restore_position() call if cursor relative position is BTR_PCUR_ON and the cursor points to the record with NOT the same field values as in a stored record(and some other not important for this case conditions). It was done because btr_pcur_restore_position() sets page_cur_mode_t mode to PAGE_CUR_LE for cursor->rel_pos == BTR_PCUR_ON before opening cursor. So we are searching for the record less or equal to stored one. And if the found record is not equal to stored one, then it is less and we need to move cursor forward. But there can be a situation when the stored record was purged, but the new one with the same key but different value was inserted while row_search_mvcc() was suspended. In this case, when the thread is awaken, it will invoke sel_restore_position_for_mysql(), which, in turns, invoke btr_pcur_restore_position(), which will return false because found record don't match stored record, and sel_restore_position_for_mysql() will move forward cursor position. The above can lead to the case when awaken row_search_mvcc() do not see records inserted by other transactions while it slept. The mtr test case shows the example how it can be. The fix is to return special value from persistent cursor restoring function which would notify its caller that uniq fields of restored record and stored record are the same, and in this case sel_restore_position_for_mysql() don't move cursor forward. Delete-marked records are correctly processed in row_search_mvcc(). Non-unique secondary indexes are "uniquified" by adding the PK, the index->n_uniq should then be index->n_fields. So there is no need in additional checks in the fix. If transaction's readview can't see the changes made in secondary index record, it requests clustered index record in row_search_mvcc() to check its transaction id and get the correspondent record version. After this row_search_mvcc() commits mtr to preserve clustered index latching order, and starts mtr. Between those mtr commit and start secondary index pages are unlatched, and purge has the ability to remove stored in the cursor record, what causes rows duplication in result set for non-locking reads, as cursor position is restored to the previously visited record. To solve this the changes are just switched off for non-locking reads, it's quite simple solution, besides the changes don't make sense for non-locking reads. The more complex and effective from performance perspective solution is to create mtr savepoint before clustered record requesting and rolling back to that savepoint after that. See MDEV-27557. One more solution is to have per-record transaction id for secondary indexes. See MDEV-17598. If any of those is implemented, just remove select_lock_type argument in sel_restore_position_for_mysql().
79 lines
3.3 KiB
Text
79 lines
3.3 KiB
Text
--source include/have_innodb.inc
|
|
--source include/count_sessions.inc
|
|
source include/have_debug.inc;
|
|
source include/have_debug_sync.inc;
|
|
|
|
CREATE TABLE t (a int PRIMARY KEY, b int NOT NULL UNIQUE) engine = InnoDB;
|
|
|
|
--connect(prevent_purge,localhost,root,,)
|
|
start transaction with consistent snapshot;
|
|
|
|
--connect(con_del_1,localhost,root,,)
|
|
INSERT INTO t VALUES (20,20);
|
|
SET DEBUG_SYNC = 'innodb_row_search_for_mysql_exit SIGNAL first_del_row_search_mvcc_finished WAIT_FOR first_del_cont';
|
|
--send DELETE FROM t WHERE b = 20
|
|
|
|
--connect(con_ins_1,localhost,root,,)
|
|
SET DEBUG_SYNC = 'now WAIT_FOR first_del_row_search_mvcc_finished';
|
|
# It's supposed the following INSERT will be suspended just after
|
|
# lock_wait_suspend_thread_enter syncpoint, and will be awaken
|
|
# after the previous DELETE commits. ib_after_row_insert will be executed
|
|
# after the INSERT is woken up. The previous DELETE will wait for
|
|
# first_del_cont signal before commit, and this signal will be sent later.
|
|
# So it's safe to use two signals in a row here, it's guaranted the first
|
|
# signal will be received before the second signal is sent.
|
|
SET DEBUG_SYNC = 'lock_wait_suspend_thread_enter SIGNAL first_ins_locked';
|
|
SET DEBUG_SYNC = 'ib_after_row_insert SIGNAL first_ins_row_inserted WAIT_FOR first_ins_cont';
|
|
--send INSERT INTO t VALUES(10, 20)
|
|
|
|
--connect(con_del_2,localhost,root,,)
|
|
SET DEBUG_SYNC = 'now WAIT_FOR first_ins_locked';
|
|
SET DEBUG_SYNC = 'lock_wait_suspend_thread_enter SIGNAL second_del_locked';
|
|
###############################################################################
|
|
# This DELETE is locked by the previous DELETE, after that DELETE is
|
|
# committed, it will still be locked by the next INSERT on delete-marked
|
|
# heap_no 2 record. After that INSERT inserted the record with heap_no 3,
|
|
# and after heap_no 2 record is purged, this DELETE will be unlocked and
|
|
# must restore persistent cursor position at heap_no 3 record, as it has the
|
|
# same secondary key value as former heap_no 2 record. Then it must be blocked
|
|
# by the previous INSERT, and after the INSERT is committed, it must
|
|
# delete the record, inserted by the previous INSERT, and the last INSERT(see
|
|
# below) must be finished without error. But instead this DELETE restores
|
|
# persistent cursor position to supremum, as a result, it does not delete the
|
|
# record, inserted by the previous INSERT, and the last INSERT is finished with
|
|
# duplicate key check error.
|
|
###############################################################################
|
|
--send DELETE FROM t WHERE b = 20
|
|
|
|
--connection default
|
|
SET DEBUG_SYNC = 'now WAIT_FOR second_del_locked';
|
|
SET DEBUG_SYNC = 'now SIGNAL first_del_cont';
|
|
SET DEBUG_SYNC = 'now WAIT_FOR first_ins_row_inserted';
|
|
--connection con_del_1
|
|
--reap
|
|
|
|
--connection default
|
|
--disconnect prevent_purge
|
|
--source include/wait_all_purged.inc
|
|
SET DEBUG_SYNC = 'now SIGNAL first_ins_cont';
|
|
|
|
--connection con_del_2
|
|
--reap
|
|
|
|
--connection con_ins_1
|
|
--reap
|
|
|
|
--connection default
|
|
###############################################################################
|
|
# Duplicate key error is expected if the bug is not fixed.
|
|
###############################################################################
|
|
INSERT INTO t VALUES(30, 20);
|
|
|
|
--disconnect con_ins_1
|
|
--disconnect con_del_1
|
|
--disconnect con_del_2
|
|
--connection default
|
|
|
|
SET DEBUG_SYNC = 'RESET';
|
|
DROP TABLE t;
|
|
--source include/wait_until_count_sessions.inc
|