Files
percona-toolkit/t/pt-table-checksum/issue_47.t
Sveta Smirnova a7efd8c94e PT-2138 - fix tests for pt table checksum (#568)
* PT-2138 - Fix tests for pt-table-checksum

- Updated t/pt-table-checksum/samples/default-results-8.0.txt and
t/pt-table-checksum/samples/static-chunk-size-results-8.0.txt
to support latest MySQL 8.0 version. Tests are now incompatible with elder 8.0 releases.
- Put fix for PT-136 into package RowChecksum
- Added execution bit for pt-online-schema-change

* PT-2138 - Fix tests for pt-table-checksum

- Fixed percona_test.load_data, so it really tests if LOAD DATA LOCAL INFILE is enabled
- Fixed option --[no]create-replicate-table, broken by commit c9836d5962

* PT-2138 - Fix tests for pt-table-checksum

- Enabled t/pt-table-checksum/error_handling.t for MySQL 8.0
- Fixed test t/pt-table-checksum/fnv_64.t and it's samples file
t/pt-table-checksum/samples/fnv64-sakila-city.txt
to reflect new function name convention and
changes after 62d84e5dba

* PT-2138 - Fix tests for pt-table-checksum

- Fixed t/pt-table-checksum/issue_1485195.t, so it checks only one table and
isn't get broken when we add tables into percona_test database
- Fixed typo in error output of bin/pt-table-checksum
- Skipped issue_47.t in 8.0 until https://jira.percona.com/browse/PT-1805 is fixed

* PT-2138 - Fix tests for pt-table-checksum

- Disabled pt-131.t for 8.0, because it does not have the QUERY_RESPONSE_TIME plugins
- Added SLOW_TESTS check to pt-1616.t
- Updated pt-226.t to include the fix for PT-1766
- For replication_filters.t: excluded false positive expression for tests 10 and 11 and added sys schema to the list of checked databases for 8.0
- Changed get_slaves in lib/MasterSlave.pm, so it returns slave's parent, required for wait_for_slaves in pt-table-checksum to work properly with chained slaves

* PT-2138 - Fix tests for pt-table-checksum

- Modified pt-204.t  to support 8.0 and diffs in system tables due to timestamps
Moved the fix for PT-1616 into the proper place: lib/NibbleIterator.pm

* PT-2138 - Fix tests for pt-table-checksum

- pxc.t -added mysql.proxies_priv into ignore list, because its timestamp is different on node
- pxc.t - removed FORK=pxc from statup options for slave (non-cluster) nodes
- pxc.t - disabled wsrep replication with help of the variable wsrep_on: sql_log_bin doesn't disable wsrep replication anymore. See https://jira.percona.com/browse/PXC-3464 for details
- Removed data.tar.gz from 5.7 sandbox configuration, because it has an outdated definition for Performance Schema
- Disabled pxc.t for version 8.0 until PT-1699 is fixed
- start-sandbox - removed the first line (ALTER USER) from the init file, because it was rewritten by the next echo command, and then repeated later.

* PT-2138 - Fix tests for pt-table-checksum

- Adopted issue_1485195.t and basics.t for MyRocks-enabled setup
- replication_filters.t - added sys schema to the list of expected schemas for 5.7 and 8.0
- issue_1485195.t - added checks for the existence of mysql.plugin, func, and proxies_priv tables
- added samples/pt-131-wipe.sql that uninstalls QRT plugin if it was earlier installed by this test
-adjusted return code in pt-204.t, because expected differences in mysql.proxies_priv

* Update lib/PerconaTest.pm

removed diagnostic code

Co-authored-by: Carlos Salguero <carlos.salguero@percona.com>

Co-authored-by: Carlos Salguero <carlos.salguero@percona.com>
2022-12-28 23:09:13 +03:00

76 lines
2.6 KiB
Perl

#!/usr/bin/env perl
BEGIN {
die "The PERCONA_TOOLKIT_BRANCH environment variable is not set.\n"
unless $ENV{PERCONA_TOOLKIT_BRANCH} && -d $ENV{PERCONA_TOOLKIT_BRANCH};
unshift @INC, "$ENV{PERCONA_TOOLKIT_BRANCH}/lib";
};
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use Test::More;
use PerconaTest;
use Sandbox;
require "$trunk/bin/pt-table-checksum";
my $dp = new DSNParser(opts=>$dsn_opts);
my $sb = new Sandbox(basedir => '/tmp', DSNParser => $dp);
my $master_dbh = $sb->get_dbh_for('master');
if ( !$master_dbh ) {
plan skip_all => 'Cannot connect to sandbox master';
} elsif ($sandbox_version >= '8.0') {
plan skip_all => "8.0 requires fix for https://jira.percona.com/browse/PT-1805";
} else {
plan tests => 3;
}
# The sandbox servers run with lock_wait_timeout=3 and it's not dynamic
# so we need to specify --set-vars innodb_lock_wait_timeout=3 else the tool will die.
# And --max-load "" prevents waiting for status variables.
my $master_dsn = 'h=127.1,P=12345,u=msandbox,p=msandbox';
my @args = ($master_dsn, qw(--set-vars innodb_lock_wait_timeout=3), '--max-load', '');
my $output;
my $row;
$sb->create_dbs($master_dbh, [qw(test)]);
$sb->load_file('master', 't/pt-table-checksum/samples/issue_47.sql');
# #############################################################################
# Issue 47: TableChunker::range_num broken for very large bigint
# #############################################################################
# pt-table-checksum 2.0 doesn't use TableChunker; it uses NibbleIterator.
# But we'll test this anyway to make sure that NibbleIterator can't handle
# very larger integers.
$output = pt_table_checksum::main(@args, qw(-t test.issue_47),
qw(--chunk-time 0 --chunk-size 3 --quiet));
is(
$output,
"0",
"No error nibbling very large int"
);
$row = $master_dbh->selectall_arrayref("select lower_boundary, upper_boundary from percona.checksums where db='test' and tbl='issue_47' order by chunk");
is_deeply(
$row,
[
[ '1', '300' ],
[ '1000', '2220293' ],
[ '65553510', '18446744073709551615' ],
[ undef, '1' ], # lower oob
[ '18446744073709551615', undef ], # upper oob
],
"Uses very large int as chunk boundary"
);
# #############################################################################
# Done.
# #############################################################################
$sb->wipe_clean($master_dbh);
ok($sb->ok(), "Sandbox servers") or BAIL_OUT(__FILE__ . " broke the sandbox");
exit;