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
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
|
# ==== Purpose ====
# Prove successful server recovery after simulated crash while
# there were general DDL queries in processing.
# The test is destined for eventual merge with binlog_crash_safe_ddl.
#
# ==== Implementation ====
# Closely follows @c see binlog_crash_safe_ddl.test.
#
# Unlike the referred (mostly) acl specific file this one provides merely
# "stub" checks after some of simulated crashes for cases that are not fully
# 2pc-ready.
# TODO: WL#9536 to refine, complete or reuse
# the stub checks.
#
# ==== References ====
# WL#9175 Correct recovery of DDL statements/transactions by binary log
# WL#9536 is responsible for adding to 2pc support to some
# of DDL:s. The current testing skips the crash-recovery testing in those case
# with tagging them as todo objects for a host WL.
#
# ==== Exempts ====
# The following DDL commands are excluded from testing:
#
# SERVER (Create etc)
# An excerpt from Docs: In MySQL 5.7, 'CREATE SERVER' is not written to the binary log,
# regardless of the logging format that is in use.
#
# LOGFILE-GROUP (Create etc)
# Not supported by Innodb.
#
# ALTER TABLESPACE
# Not supported by Innodb.
#
# === Limitations ===
# --force-restart is required to run the test.
#
# When the object deletion was interrupted by crash
# there could be errors on object recreation attempt like
# 1007: Can't create database '...'; database exists
# due to
# Bug #25663287 DROP DATABASE DOESN'T GETS RECOVERED CORRECTLY UPON POST BINLOG CRASH/RESTART
# or asserts like in
# Bug #25651042 TANSACTION DOES NOT POST-CRASH RECOVER IN INNODB EVEN IF PREPARED AND BINLOGGED
#
--source include/not_valgrind.inc
--source include/not_crashrep.inc
--source include/have_log_bin.inc
--source include/have_debug.inc
--source include/have_binlog_format_mixed.inc
--disable_query_log
call mtr.add_suppression("Could not find a valid tablespace file for");
--enable_query_log
# The *internal* var is used by gtid_step_assert.inc
--let $gtid_mode_on= `SELECT @@GLOBAL.GTID_MODE = 'ON'`
# To use by gtid_assert
--let $gtid_mode= `SELECT @@GLOBAL.GTID_MODE`
--source include/gtid_utils.inc
# Instructing gtid_step_assert called by sourced file
# to work in gtid-mode OFF as well.
# Besides, there's a reason for a bug to fix 'cos setting $gtid_mode
# turns out insufficient for $gtid_step_assert to keep the value 1 for
# an internal "_"-var of include/gtid_step_assert.inc.
--let $gtid_step_gtid_mode_agnostic=`SELECT '$gtid_mode' != 'ON'`
# The post_prepare crash is simulated only when the correspoding feature
# is fully 2pc-supported. The default set of simulation control parameters
# below assumes a feature can't be rolled back, but can be committed
# once found prepared and logged.
--let $do_pre_binlog=0
--let $do_post_binlog=1
--let $do_only_regular_logging=0
# todo: the $pre_binlog_crash_check:s are to be activated/refined once
# the corresponding has learned how to rollback (WL9536).
# E.g CREATE|DROP|ALTER|RENAME table.
--let $table=t_2
--source extra/binlog_tests/binlog_ddl_half_atomic.inc
# Cleanup.
--source include/gtid_utils_end.inc
|
