File: zlob_big_rollback.test

package info (click to toggle)
mysql-8.0 8.0.43-3
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 1,273,924 kB
  • sloc: cpp: 4,684,605; ansic: 412,450; pascal: 108,398; java: 83,641; perl: 30,221; cs: 27,067; sql: 26,594; sh: 24,181; python: 21,816; yacc: 17,169; php: 11,522; xml: 7,388; javascript: 7,076; makefile: 2,194; lex: 1,075; awk: 670; asm: 520; objc: 183; ruby: 97; lisp: 86
file content (22 lines) | stat: -rw-r--r-- 796 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
#
# Bug #29846292 ROLLBACK OF BIG TRANSACTION DUE TO CONFLICT REMAINS IN HUNG STATE
#
--source include/have_innodb_max_16k.inc

# In this scenario, we create a LOB which does not fit in the buffer pool, thus
# the rollback has to be smart about managing pages.
# In particular a buggy implementaiton which tries to keep all pages of the LOB
# in the memory until mini-transaction commits, will deadlock with itself at
# some point when trying to fetch next page of the LOB from storage to already
# fully saturated buffer pool.

SET GLOBAL innodb_compression_level = 0;

CREATE TABLE t1 (id INT PRIMARY KEY NOT NULL, val LONGTEXT)
ENGINE=InnoDB ROW_FORMAT=compressed;
BEGIN;
INSERT INTO t1 VALUES (1, repeat('a',6000000));
ROLLBACK;
DROP TABLE t1;

SET GLOBAL innodb_compression_level = default;