1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
|
#
# Bug #23070734 CONCURRENT TRUNCATE TABLES CAUSE STALLS
#
create table t1 (f1 int ,f2 int,key(f2)) engine=innodb;
begin;
insert into t1 values (10,45),(20,78),(30,88),(40,23),(50,78),(60,11),(70,56),(80,90);
delete from t1;
connect con2,localhost,root,,;
START TRANSACTION WITH CONSISTENT SNAPSHOT;
connection default;
SET @saved_frequency = @@GLOBAL.innodb_purge_rseg_truncate_frequency;
SET GLOBAL innodb_purge_rseg_truncate_frequency = 1;
commit;
connect con1,localhost,root,,;
SET DEBUG_SYNC= 'buffer_pool_scan SIGNAL started WAIT_FOR finish_scan';
truncate table t1;
connection con2;
SET DEBUG_SYNC= 'now WAIT_FOR started';
COMMIT;
disconnect con2;
connection default;
InnoDB 0 transactions not purged
SET GLOBAL innodb_purge_rseg_truncate_frequency = @saved_frequency;
SET DEBUG_SYNC = 'now SIGNAL finish_scan';
SET DEBUG_SYNC = 'RESET';
connection con1;
disconnect con1;
connection default;
drop table t1;
|
