| 1 |
--source include/have_debug_sync.inc |
| 2 |
|
| 3 |
--echo # Bug #28176910 |
| 4 |
--echo # PERFORMANCE_SCHEMA.DATA_LOCKS.OBJECT_INSTANCE_BEGIN |
| 5 |
--echo # CHANGES DEPENDING ON ORDER |
| 6 |
|
| 7 |
# This scenario is meant to create a situation in which a single |
| 8 |
# transaction has two locks for the same record, which differ only |
| 9 |
# in their mode (X vs X,REC_NOT_GAP) which historicaly caused a bug |
| 10 |
# in that both of them had the same performance.data_locks.engine_lock_id. |
| 11 |
# Having two rows with the same engline_lock_id caused visible problems: |
| 12 |
# using SELECT...FROM data_locks ORDER BY ... caused one of the two rows |
| 13 |
# to be reported twice and the other row not even once. |
| 14 |
# There were also problems with data_lock_waits which should contain |
| 15 |
# exactly one edge for each pair of conflicting locks, however having |
| 16 |
# two rows with the same id in data_locks could lead to strange results |
| 17 |
# when joining data_locks with data_lock_waits. |
| 18 |
# Finally, the information_schema.innodb_trx.trx_requested_lock_id could |
| 19 |
# be joined with performance_schema.data_locks.engine_lock_id again |
| 20 |
# leading to strange results. |
| 21 |
# This test checks that: |
| 22 |
# - engline_lock_id values are unique |
| 23 |
# - that ORDER BY does not lead to duplicates |
| 24 |
# - that JOIN with data_lock_waits works correctly |
| 25 |
# - that JOIN with innodb_trx works correctly |
| 26 |
|
| 27 |
CREATE TABLE t1( |
| 28 |
id INT PRIMARY KEY |
| 29 |
) Engine=InnoDB; |
| 30 |
|
| 31 |
INSERT INTO t1 VALUES (1); |
| 32 |
# Create a situation in which a transaction holds more than one lock on same record. |
| 33 |
BEGIN; |
| 34 |
SELECT * FROM t1 WHERE id=1 FOR SHARE; |
| 35 |
SELECT * FROM t1 WHERE id=1 FOR UPDATE; |
| 36 |
|
| 37 |
SELECT COUNT(DISTINCT engine_lock_id) FROM performance_schema.data_locks; |
| 38 |
|
| 39 |
SELECT |
| 40 |
lock_type, |
| 41 |
lock_mode, |
| 42 |
lock_status, |
| 43 |
lock_data |
| 44 |
FROM performance_schema.data_locks |
| 45 |
ORDER BY 1,2,3,4; |
| 46 |
|
| 47 |
--connect (C1, localhost, root,,) |
| 48 |
|
| 49 |
BEGIN; |
| 50 |
SET DEBUG_SYNC = 'lock_wait_will_wait SIGNAL C1_will_wait'; |
| 51 |
--send SELECT * FROM t1 FOR UPDATE; |
| 52 |
|
| 53 |
--connection default |
| 54 |
SET DEBUG_SYNC = 'now WAIT_FOR C1_will_wait'; |
| 55 |
SELECT |
| 56 |
r.object_name,r.lock_type,r.lock_mode,r.lock_status,r.lock_data , |
| 57 |
b.object_name,b.lock_type,b.lock_mode,b.lock_status,b.lock_data |
| 58 |
FROM performance_schema.data_lock_waits |
| 59 |
JOIN performance_schema.data_locks r |
| 60 |
ON(r.engine_lock_id = requesting_engine_lock_id) |
| 61 |
JOIN performance_schema.data_locks b |
| 62 |
ON(b.engine_lock_id = blocking_engine_lock_id) |
| 63 |
ORDER BY b.lock_mode; |
| 64 |
|
| 65 |
SELECT r.object_name,r.lock_type,r.lock_mode,r.lock_status,r.lock_data |
| 66 |
FROM information_schema.innodb_trx |
| 67 |
JOIN performance_schema.data_locks r |
| 68 |
ON(r.engine_lock_id = trx_requested_lock_id); |
| 69 |
|
| 70 |
|
| 71 |
ROLLBACK; |
| 72 |
|
| 73 |
--connection C1 |
| 74 |
--reap |
| 75 |
ROLLBACK; |
| 76 |
|
| 77 |
--connection default |
| 78 |
--disconnect C1 |
| 79 |
|
| 80 |
DROP TABLE t1; |