system: Linux mars.sprixweb.com 3.10.0-1160.119.1.el7.x86_64 #1 SMP Tue Jun 4 14:43:51 UTC 2024 x86_64
| Direktori : /bin/ |
|
|
| Current File : //bin/pt-duplicate-key-checker |
#!/usr/bin/env perl
# This program is part of Percona Toolkit: http://www.percona.com/software/
# See "COPYRIGHT, LICENSE, AND WARRANTY" at the end of this file for legal
# notices and disclaimers.
use strict;
use warnings FATAL => 'all';
# This tool is "fat-packed": most of its dependent modules are embedded
# in this file. Setting %INC to this file for each module makes Perl aware
# of this so it will not try to load the module from @INC. See the tool's
# documentation for a full list of dependencies.
BEGIN {
$INC{$_} = __FILE__ for map { (my $pkg = "$_.pm") =~ s!::!/!g; $pkg } (qw(
Percona::Toolkit
Quoter
TableParser
DSNParser
OptionParser
KeySize
DuplicateKeyFinder
Daemon
Schema
SchemaIterator
HTTP::Micro
VersionCheck
));
}
# ###########################################################################
# Percona::Toolkit package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/Percona/Toolkit.pm
# t/lib/Percona/Toolkit.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package Percona::Toolkit;
our $VERSION = '2.2.15';
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use Carp qw(carp cluck);
use Data::Dumper qw();
require Exporter;
our @ISA = qw(Exporter);
our @EXPORT_OK = qw(
have_required_args
Dumper
_d
);
sub have_required_args {
my ($args, @required_args) = @_;
my $have_required_args = 1;
foreach my $arg ( @required_args ) {
if ( !defined $args->{$arg} ) {
$have_required_args = 0;
carp "Argument $arg is not defined";
}
}
cluck unless $have_required_args; # print backtrace
return $have_required_args;
}
sub Dumper {
local $Data::Dumper::Indent = 1;
local $Data::Dumper::Sortkeys = 1;
local $Data::Dumper::Quotekeys = 0;
Data::Dumper::Dumper(@_);
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End Percona::Toolkit package
# ###########################################################################
# ###########################################################################
# Quoter package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/Quoter.pm
# t/lib/Quoter.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package Quoter;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use Data::Dumper;
$Data::Dumper::Indent = 1;
$Data::Dumper::Sortkeys = 1;
$Data::Dumper::Quotekeys = 0;
sub new {
my ( $class, %args ) = @_;
return bless {}, $class;
}
sub quote {
my ( $self, @vals ) = @_;
foreach my $val ( @vals ) {
$val =~ s/`/``/g;
}
return join('.', map { '`' . $_ . '`' } @vals);
}
sub quote_val {
my ( $self, $val, %args ) = @_;
return 'NULL' unless defined $val; # undef = NULL
return "''" if $val eq ''; # blank string = ''
return $val if $val =~ m/^0x[0-9a-fA-F]+$/ # quote hex data
&& !$args{is_char}; # unless is_char is true
return $val if $args{is_float};
$val =~ s/(['\\])/\\$1/g;
return "'$val'";
}
sub split_unquote {
my ( $self, $db_tbl, $default_db ) = @_;
my ( $db, $tbl ) = split(/[.]/, $db_tbl);
if ( !$tbl ) {
$tbl = $db;
$db = $default_db;
}
for ($db, $tbl) {
next unless $_;
s/\A`//;
s/`\z//;
s/``/`/g;
}
return ($db, $tbl);
}
sub literal_like {
my ( $self, $like ) = @_;
return unless $like;
$like =~ s/([%_])/\\$1/g;
return "'$like'";
}
sub join_quote {
my ( $self, $default_db, $db_tbl ) = @_;
return unless $db_tbl;
my ($db, $tbl) = split(/[.]/, $db_tbl);
if ( !$tbl ) {
$tbl = $db;
$db = $default_db;
}
$db = "`$db`" if $db && $db !~ m/^`/;
$tbl = "`$tbl`" if $tbl && $tbl !~ m/^`/;
return $db ? "$db.$tbl" : $tbl;
}
sub serialize_list {
my ( $self, @args ) = @_;
PTDEBUG && _d('Serializing', Dumper(\@args));
return unless @args;
my @parts;
foreach my $arg ( @args ) {
if ( defined $arg ) {
$arg =~ s/,/\\,/g; # escape commas
$arg =~ s/\\N/\\\\N/g; # escape literal \N
push @parts, $arg;
}
else {
push @parts, '\N';
}
}
my $string = join(',', @parts);
PTDEBUG && _d('Serialized: <', $string, '>');
return $string;
}
sub deserialize_list {
my ( $self, $string ) = @_;
PTDEBUG && _d('Deserializing <', $string, '>');
die "Cannot deserialize an undefined string" unless defined $string;
my @parts;
foreach my $arg ( split(/(?<!\\),/, $string) ) {
if ( $arg eq '\N' ) {
$arg = undef;
}
else {
$arg =~ s/\\,/,/g;
$arg =~ s/\\\\N/\\N/g;
}
push @parts, $arg;
}
if ( !@parts ) {
my $n_empty_strings = $string =~ tr/,//;
$n_empty_strings++;
PTDEBUG && _d($n_empty_strings, 'empty strings');
map { push @parts, '' } 1..$n_empty_strings;
}
elsif ( $string =~ m/(?<!\\),$/ ) {
PTDEBUG && _d('Last value is an empty string');
push @parts, '';
}
PTDEBUG && _d('Deserialized', Dumper(\@parts));
return @parts;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End Quoter package
# ###########################################################################
# ###########################################################################
# TableParser package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/TableParser.pm
# t/lib/TableParser.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package TableParser;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use Data::Dumper;
$Data::Dumper::Indent = 1;
$Data::Dumper::Sortkeys = 1;
$Data::Dumper::Quotekeys = 0;
local $EVAL_ERROR;
eval {
require Quoter;
};
sub new {
my ( $class, %args ) = @_;
my $self = { %args };
$self->{Quoter} ||= Quoter->new();
return bless $self, $class;
}
sub Quoter { shift->{Quoter} }
sub get_create_table {
my ( $self, $dbh, $db, $tbl ) = @_;
die "I need a dbh parameter" unless $dbh;
die "I need a db parameter" unless $db;
die "I need a tbl parameter" unless $tbl;
my $q = $self->{Quoter};
my $new_sql_mode
= q{/*!40101 SET @OLD_SQL_MODE := @@SQL_MODE, }
. q{@@SQL_MODE := '', }
. q{@OLD_QUOTE := @@SQL_QUOTE_SHOW_CREATE, }
. q{@@SQL_QUOTE_SHOW_CREATE := 1 */};
my $old_sql_mode
= q{/*!40101 SET @@SQL_MODE := @OLD_SQL_MODE, }
. q{@@SQL_QUOTE_SHOW_CREATE := @OLD_QUOTE */};
PTDEBUG && _d($new_sql_mode);
eval { $dbh->do($new_sql_mode); };
PTDEBUG && $EVAL_ERROR && _d($EVAL_ERROR);
my $use_sql = 'USE ' . $q->quote($db);
PTDEBUG && _d($dbh, $use_sql);
$dbh->do($use_sql);
my $show_sql = "SHOW CREATE TABLE " . $q->quote($db, $tbl);
PTDEBUG && _d($show_sql);
my $href;
eval { $href = $dbh->selectrow_hashref($show_sql); };
if ( my $e = $EVAL_ERROR ) {
PTDEBUG && _d($old_sql_mode);
$dbh->do($old_sql_mode);
die $e;
}
PTDEBUG && _d($old_sql_mode);
$dbh->do($old_sql_mode);
my ($key) = grep { m/create (?:table|view)/i } keys %$href;
if ( !$key ) {
die "Error: no 'Create Table' or 'Create View' in result set from "
. "$show_sql: " . Dumper($href);
}
return $href->{$key};
}
sub parse {
my ( $self, $ddl, $opts ) = @_;
return unless $ddl;
if ( $ddl =~ m/CREATE (?:TEMPORARY )?TABLE "/ ) {
$ddl = $self->ansi_to_legacy($ddl);
}
elsif ( $ddl !~ m/CREATE (?:TEMPORARY )?TABLE `/ ) {
die "TableParser doesn't handle CREATE TABLE without quoting.";
}
my ($name) = $ddl =~ m/CREATE (?:TEMPORARY )?TABLE\s+(`.+?`)/;
(undef, $name) = $self->{Quoter}->split_unquote($name) if $name;
$ddl =~ s/(`[^`]+`)/\L$1/g;
my $engine = $self->get_engine($ddl);
my @defs = $ddl =~ m/^(\s+`.*?),?$/gm;
my @cols = map { $_ =~ m/`([^`]+)`/ } @defs;
PTDEBUG && _d('Table cols:', join(', ', map { "`$_`" } @cols));
my %def_for;
@def_for{@cols} = @defs;
my (@nums, @null);
my (%type_for, %is_nullable, %is_numeric, %is_autoinc);
foreach my $col ( @cols ) {
my $def = $def_for{$col};
$def =~ s/``//g;
my ( $type ) = $def =~ m/`[^`]+`\s([a-z]+)/;
die "Can't determine column type for $def" unless $type;
$type_for{$col} = $type;
if ( $type =~ m/(?:(?:tiny|big|medium|small)?int|float|double|decimal|year)/ ) {
push @nums, $col;
$is_numeric{$col} = 1;
}
if ( $def !~ m/NOT NULL/ ) {
push @null, $col;
$is_nullable{$col} = 1;
}
$is_autoinc{$col} = $def =~ m/AUTO_INCREMENT/i ? 1 : 0;
}
my ($keys, $clustered_key) = $self->get_keys($ddl, $opts, \%is_nullable);
my ($charset) = $ddl =~ m/DEFAULT CHARSET=(\w+)/;
return {
name => $name,
cols => \@cols,
col_posn => { map { $cols[$_] => $_ } 0..$#cols },
is_col => { map { $_ => 1 } @cols },
null_cols => \@null,
is_nullable => \%is_nullable,
is_autoinc => \%is_autoinc,
clustered_key => $clustered_key,
keys => $keys,
defs => \%def_for,
numeric_cols => \@nums,
is_numeric => \%is_numeric,
engine => $engine,
type_for => \%type_for,
charset => $charset,
};
}
sub sort_indexes {
my ( $self, $tbl ) = @_;
my @indexes
= sort {
(($a ne 'PRIMARY') <=> ($b ne 'PRIMARY'))
|| ( !$tbl->{keys}->{$a}->{is_unique} <=> !$tbl->{keys}->{$b}->{is_unique} )
|| ( $tbl->{keys}->{$a}->{is_nullable} <=> $tbl->{keys}->{$b}->{is_nullable} )
|| ( scalar(@{$tbl->{keys}->{$a}->{cols}}) <=> scalar(@{$tbl->{keys}->{$b}->{cols}}) )
}
grep {
$tbl->{keys}->{$_}->{type} eq 'BTREE'
}
sort keys %{$tbl->{keys}};
PTDEBUG && _d('Indexes sorted best-first:', join(', ', @indexes));
return @indexes;
}
sub find_best_index {
my ( $self, $tbl, $index ) = @_;
my $best;
if ( $index ) {
($best) = grep { uc $_ eq uc $index } keys %{$tbl->{keys}};
}
if ( !$best ) {
if ( $index ) {
die "Index '$index' does not exist in table";
}
else {
($best) = $self->sort_indexes($tbl);
}
}
PTDEBUG && _d('Best index found is', $best);
return $best;
}
sub find_possible_keys {
my ( $self, $dbh, $database, $table, $quoter, $where ) = @_;
return () unless $where;
my $sql = 'EXPLAIN SELECT * FROM ' . $quoter->quote($database, $table)
. ' WHERE ' . $where;
PTDEBUG && _d($sql);
my $expl = $dbh->selectrow_hashref($sql);
$expl = { map { lc($_) => $expl->{$_} } keys %$expl };
if ( $expl->{possible_keys} ) {
PTDEBUG && _d('possible_keys =', $expl->{possible_keys});
my @candidates = split(',', $expl->{possible_keys});
my %possible = map { $_ => 1 } @candidates;
if ( $expl->{key} ) {
PTDEBUG && _d('MySQL chose', $expl->{key});
unshift @candidates, grep { $possible{$_} } split(',', $expl->{key});
PTDEBUG && _d('Before deduping:', join(', ', @candidates));
my %seen;
@candidates = grep { !$seen{$_}++ } @candidates;
}
PTDEBUG && _d('Final list:', join(', ', @candidates));
return @candidates;
}
else {
PTDEBUG && _d('No keys in possible_keys');
return ();
}
}
sub check_table {
my ( $self, %args ) = @_;
my @required_args = qw(dbh db tbl);
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
}
my ($dbh, $db, $tbl) = @args{@required_args};
my $q = $self->{Quoter} || 'Quoter';
my $db_tbl = $q->quote($db, $tbl);
PTDEBUG && _d('Checking', $db_tbl);
my $sql = "SHOW TABLES FROM " . $q->quote($db)
. ' LIKE ' . $q->literal_like($tbl);
PTDEBUG && _d($sql);
my $row;
eval {
$row = $dbh->selectrow_arrayref($sql);
};
if ( $EVAL_ERROR ) {
PTDEBUG && _d($EVAL_ERROR);
return 0;
}
if ( !$row->[0] || $row->[0] ne $tbl ) {
PTDEBUG && _d('Table does not exist');
return 0;
}
PTDEBUG && _d('Table', $db, $tbl, 'exists');
return 1;
}
sub get_engine {
my ( $self, $ddl, $opts ) = @_;
my ( $engine ) = $ddl =~ m/\).*?(?:ENGINE|TYPE)=(\w+)/;
PTDEBUG && _d('Storage engine:', $engine);
return $engine || undef;
}
sub get_keys {
my ( $self, $ddl, $opts, $is_nullable ) = @_;
my $engine = $self->get_engine($ddl);
my $keys = {};
my $clustered_key = undef;
KEY:
foreach my $key ( $ddl =~ m/^ ((?:[A-Z]+ )?KEY .*)$/gm ) {
next KEY if $key =~ m/FOREIGN/;
my $key_ddl = $key;
PTDEBUG && _d('Parsed key:', $key_ddl);
if ( !$engine || $engine !~ m/MEMORY|HEAP/ ) {
$key =~ s/USING HASH/USING BTREE/;
}
my ( $type, $cols ) = $key =~ m/(?:USING (\w+))? \((.+)\)/;
my ( $special ) = $key =~ m/(FULLTEXT|SPATIAL)/;
$type = $type || $special || 'BTREE';
my ($name) = $key =~ m/(PRIMARY|`[^`]*`)/;
my $unique = $key =~ m/PRIMARY|UNIQUE/ ? 1 : 0;
my @cols;
my @col_prefixes;
foreach my $col_def ( $cols =~ m/`[^`]+`(?:\(\d+\))?/g ) {
my ($name, $prefix) = $col_def =~ m/`([^`]+)`(?:\((\d+)\))?/;
push @cols, $name;
push @col_prefixes, $prefix;
}
$name =~ s/`//g;
PTDEBUG && _d( $name, 'key cols:', join(', ', map { "`$_`" } @cols));
$keys->{$name} = {
name => $name,
type => $type,
colnames => $cols,
cols => \@cols,
col_prefixes => \@col_prefixes,
is_unique => $unique,
is_nullable => scalar(grep { $is_nullable->{$_} } @cols),
is_col => { map { $_ => 1 } @cols },
ddl => $key_ddl,
};
if ( ($engine || '') =~ m/InnoDB/i && !$clustered_key ) {
my $this_key = $keys->{$name};
if ( $this_key->{name} eq 'PRIMARY' ) {
$clustered_key = 'PRIMARY';
}
elsif ( $this_key->{is_unique} && !$this_key->{is_nullable} ) {
$clustered_key = $this_key->{name};
}
PTDEBUG && $clustered_key && _d('This key is the clustered key');
}
}
return $keys, $clustered_key;
}
sub get_fks {
my ( $self, $ddl, $opts ) = @_;
my $q = $self->{Quoter};
my $fks = {};
foreach my $fk (
$ddl =~ m/CONSTRAINT .* FOREIGN KEY .* REFERENCES [^\)]*\)/mg )
{
my ( $name ) = $fk =~ m/CONSTRAINT `(.*?)`/;
my ( $cols ) = $fk =~ m/FOREIGN KEY \(([^\)]+)\)/;
my ( $parent, $parent_cols ) = $fk =~ m/REFERENCES (\S+) \(([^\)]+)\)/;
my ($db, $tbl) = $q->split_unquote($parent, $opts->{database});
my %parent_tbl = (tbl => $tbl);
$parent_tbl{db} = $db if $db;
if ( $parent !~ m/\./ && $opts->{database} ) {
$parent = $q->quote($opts->{database}) . ".$parent";
}
$fks->{$name} = {
name => $name,
colnames => $cols,
cols => [ map { s/[ `]+//g; $_; } split(',', $cols) ],
parent_tbl => \%parent_tbl,
parent_tblname => $parent,
parent_cols => [ map { s/[ `]+//g; $_; } split(',', $parent_cols) ],
parent_colnames=> $parent_cols,
ddl => $fk,
};
}
return $fks;
}
sub remove_auto_increment {
my ( $self, $ddl ) = @_;
$ddl =~ s/(^\).*?) AUTO_INCREMENT=\d+\b/$1/m;
return $ddl;
}
sub get_table_status {
my ( $self, $dbh, $db, $like ) = @_;
my $q = $self->{Quoter};
my $sql = "SHOW TABLE STATUS FROM " . $q->quote($db);
my @params;
if ( $like ) {
$sql .= ' LIKE ?';
push @params, $like;
}
PTDEBUG && _d($sql, @params);
my $sth = $dbh->prepare($sql);
eval { $sth->execute(@params); };
if ($EVAL_ERROR) {
PTDEBUG && _d($EVAL_ERROR);
return;
}
my @tables = @{$sth->fetchall_arrayref({})};
@tables = map {
my %tbl; # Make a copy with lowercased keys
@tbl{ map { lc $_ } keys %$_ } = values %$_;
$tbl{engine} ||= $tbl{type} || $tbl{comment};
delete $tbl{type};
\%tbl;
} @tables;
return @tables;
}
my $ansi_quote_re = qr/" [^"]* (?: "" [^"]* )* (?<=.) "/ismx;
sub ansi_to_legacy {
my ($self, $ddl) = @_;
$ddl =~ s/($ansi_quote_re)/ansi_quote_replace($1)/ge;
return $ddl;
}
sub ansi_quote_replace {
my ($val) = @_;
$val =~ s/^"|"$//g;
$val =~ s/`/``/g;
$val =~ s/""/"/g;
return "`$val`";
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End TableParser package
# ###########################################################################
# ###########################################################################
# DSNParser package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/DSNParser.pm
# t/lib/DSNParser.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package DSNParser;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use Data::Dumper;
$Data::Dumper::Indent = 0;
$Data::Dumper::Quotekeys = 0;
my $dsn_sep = qr/(?<!\\),/;
eval {
require DBI;
};
my $have_dbi = $EVAL_ERROR ? 0 : 1;
sub new {
my ( $class, %args ) = @_;
foreach my $arg ( qw(opts) ) {
die "I need a $arg argument" unless $args{$arg};
}
my $self = {
opts => {} # h, P, u, etc. Should come from DSN OPTIONS section in POD.
};
foreach my $opt ( @{$args{opts}} ) {
if ( !$opt->{key} || !$opt->{desc} ) {
die "Invalid DSN option: ", Dumper($opt);
}
PTDEBUG && _d('DSN option:',
join(', ',
map { "$_=" . (defined $opt->{$_} ? ($opt->{$_} || '') : 'undef') }
keys %$opt
)
);
$self->{opts}->{$opt->{key}} = {
dsn => $opt->{dsn},
desc => $opt->{desc},
copy => $opt->{copy} || 0,
};
}
return bless $self, $class;
}
sub prop {
my ( $self, $prop, $value ) = @_;
if ( @_ > 2 ) {
PTDEBUG && _d('Setting', $prop, 'property');
$self->{$prop} = $value;
}
return $self->{$prop};
}
sub parse {
my ( $self, $dsn, $prev, $defaults ) = @_;
if ( !$dsn ) {
PTDEBUG && _d('No DSN to parse');
return;
}
PTDEBUG && _d('Parsing', $dsn);
$prev ||= {};
$defaults ||= {};
my %given_props;
my %final_props;
my $opts = $self->{opts};
foreach my $dsn_part ( split($dsn_sep, $dsn) ) {
$dsn_part =~ s/\\,/,/g;
if ( my ($prop_key, $prop_val) = $dsn_part =~ m/^(.)=(.*)$/ ) {
$given_props{$prop_key} = $prop_val;
}
else {
PTDEBUG && _d('Interpreting', $dsn_part, 'as h=', $dsn_part);
$given_props{h} = $dsn_part;
}
}
foreach my $key ( keys %$opts ) {
PTDEBUG && _d('Finding value for', $key);
$final_props{$key} = $given_props{$key};
if ( !defined $final_props{$key}
&& defined $prev->{$key} && $opts->{$key}->{copy} )
{
$final_props{$key} = $prev->{$key};
PTDEBUG && _d('Copying value for', $key, 'from previous DSN');
}
if ( !defined $final_props{$key} ) {
$final_props{$key} = $defaults->{$key};
PTDEBUG && _d('Copying value for', $key, 'from defaults');
}
}
foreach my $key ( keys %given_props ) {
die "Unknown DSN option '$key' in '$dsn'. For more details, "
. "please use the --help option, or try 'perldoc $PROGRAM_NAME' "
. "for complete documentation."
unless exists $opts->{$key};
}
if ( (my $required = $self->prop('required')) ) {
foreach my $key ( keys %$required ) {
die "Missing required DSN option '$key' in '$dsn'. For more details, "
. "please use the --help option, or try 'perldoc $PROGRAM_NAME' "
. "for complete documentation."
unless $final_props{$key};
}
}
return \%final_props;
}
sub parse_options {
my ( $self, $o ) = @_;
die 'I need an OptionParser object' unless ref $o eq 'OptionParser';
my $dsn_string
= join(',',
map { "$_=".$o->get($_); }
grep { $o->has($_) && $o->get($_) }
keys %{$self->{opts}}
);
PTDEBUG && _d('DSN string made from options:', $dsn_string);
return $self->parse($dsn_string);
}
sub as_string {
my ( $self, $dsn, $props ) = @_;
return $dsn unless ref $dsn;
my @keys = $props ? @$props : sort keys %$dsn;
return join(',',
map { "$_=" . ($_ eq 'p' ? '...' : $dsn->{$_}) }
grep {
exists $self->{opts}->{$_}
&& exists $dsn->{$_}
&& defined $dsn->{$_}
} @keys);
}
sub usage {
my ( $self ) = @_;
my $usage
= "DSN syntax is key=value[,key=value...] Allowable DSN keys:\n\n"
. " KEY COPY MEANING\n"
. " === ==== =============================================\n";
my %opts = %{$self->{opts}};
foreach my $key ( sort keys %opts ) {
$usage .= " $key "
. ($opts{$key}->{copy} ? 'yes ' : 'no ')
. ($opts{$key}->{desc} || '[No description]')
. "\n";
}
$usage .= "\n If the DSN is a bareword, the word is treated as the 'h' key.\n";
return $usage;
}
sub get_cxn_params {
my ( $self, $info ) = @_;
my $dsn;
my %opts = %{$self->{opts}};
my $driver = $self->prop('dbidriver') || '';
if ( $driver eq 'Pg' ) {
$dsn = 'DBI:Pg:dbname=' . ( $info->{D} || '' ) . ';'
. join(';', map { "$opts{$_}->{dsn}=$info->{$_}" }
grep { defined $info->{$_} }
qw(h P));
}
else {
$dsn = 'DBI:mysql:' . ( $info->{D} || '' ) . ';'
. join(';', map { "$opts{$_}->{dsn}=$info->{$_}" }
grep { defined $info->{$_} }
qw(F h P S A))
. ';mysql_read_default_group=client'
. ($info->{L} ? ';mysql_local_infile=1' : '');
}
PTDEBUG && _d($dsn);
return ($dsn, $info->{u}, $info->{p});
}
sub fill_in_dsn {
my ( $self, $dbh, $dsn ) = @_;
my $vars = $dbh->selectall_hashref('SHOW VARIABLES', 'Variable_name');
my ($user, $db) = $dbh->selectrow_array('SELECT USER(), DATABASE()');
$user =~ s/@.*//;
$dsn->{h} ||= $vars->{hostname}->{Value};
$dsn->{S} ||= $vars->{'socket'}->{Value};
$dsn->{P} ||= $vars->{port}->{Value};
$dsn->{u} ||= $user;
$dsn->{D} ||= $db;
}
sub get_dbh {
my ( $self, $cxn_string, $user, $pass, $opts ) = @_;
$opts ||= {};
my $defaults = {
AutoCommit => 0,
RaiseError => 1,
PrintError => 0,
ShowErrorStatement => 1,
mysql_enable_utf8 => ($cxn_string =~ m/charset=utf8/i ? 1 : 0),
};
@{$defaults}{ keys %$opts } = values %$opts;
if (delete $defaults->{L}) { # L for LOAD DATA LOCAL INFILE, our own extension
$defaults->{mysql_local_infile} = 1;
}
if ( $opts->{mysql_use_result} ) {
$defaults->{mysql_use_result} = 1;
}
if ( !$have_dbi ) {
die "Cannot connect to MySQL because the Perl DBI module is not "
. "installed or not found. Run 'perl -MDBI' to see the directories "
. "that Perl searches for DBI. If DBI is not installed, try:\n"
. " Debian/Ubuntu apt-get install libdbi-perl\n"
. " RHEL/CentOS yum install perl-DBI\n"
. " OpenSolaris pkg install pkg:/SUNWpmdbi\n";
}
my $dbh;
my $tries = 2;
while ( !$dbh && $tries-- ) {
PTDEBUG && _d($cxn_string, ' ', $user, ' ', $pass,
join(', ', map { "$_=>$defaults->{$_}" } keys %$defaults ));
$dbh = eval { DBI->connect($cxn_string, $user, $pass, $defaults) };
if ( !$dbh && $EVAL_ERROR ) {
if ( $EVAL_ERROR =~ m/locate DBD\/mysql/i ) {
die "Cannot connect to MySQL because the Perl DBD::mysql module is "
. "not installed or not found. Run 'perl -MDBD::mysql' to see "
. "the directories that Perl searches for DBD::mysql. If "
. "DBD::mysql is not installed, try:\n"
. " Debian/Ubuntu apt-get install libdbd-mysql-perl\n"
. " RHEL/CentOS yum install perl-DBD-MySQL\n"
. " OpenSolaris pgk install pkg:/SUNWapu13dbd-mysql\n";
}
elsif ( $EVAL_ERROR =~ m/not a compiled character set|character set utf8/ ) {
PTDEBUG && _d('Going to try again without utf8 support');
delete $defaults->{mysql_enable_utf8};
}
if ( !$tries ) {
die $EVAL_ERROR;
}
}
}
if ( $cxn_string =~ m/mysql/i ) {
my $sql;
$sql = 'SELECT @@SQL_MODE';
PTDEBUG && _d($dbh, $sql);
my ($sql_mode) = eval { $dbh->selectrow_array($sql) };
if ( $EVAL_ERROR ) {
die "Error getting the current SQL_MODE: $EVAL_ERROR";
}
if ( my ($charset) = $cxn_string =~ m/charset=([\w]+)/ ) {
$sql = qq{/*!40101 SET NAMES "$charset"*/};
PTDEBUG && _d($dbh, $sql);
eval { $dbh->do($sql) };
if ( $EVAL_ERROR ) {
die "Error setting NAMES to $charset: $EVAL_ERROR";
}
PTDEBUG && _d('Enabling charset for STDOUT');
if ( $charset eq 'utf8' ) {
binmode(STDOUT, ':utf8')
or die "Can't binmode(STDOUT, ':utf8'): $OS_ERROR";
}
else {
binmode(STDOUT) or die "Can't binmode(STDOUT): $OS_ERROR";
}
}
if ( my $vars = $self->prop('set-vars') ) {
$self->set_vars($dbh, $vars);
}
$sql = 'SET @@SQL_QUOTE_SHOW_CREATE = 1'
. '/*!40101, @@SQL_MODE=\'NO_AUTO_VALUE_ON_ZERO'
. ($sql_mode ? ",$sql_mode" : '')
. '\'*/';
PTDEBUG && _d($dbh, $sql);
eval { $dbh->do($sql) };
if ( $EVAL_ERROR ) {
die "Error setting SQL_QUOTE_SHOW_CREATE, SQL_MODE"
. ($sql_mode ? " and $sql_mode" : '')
. ": $EVAL_ERROR";
}
}
PTDEBUG && _d('DBH info: ',
$dbh,
Dumper($dbh->selectrow_hashref(
'SELECT DATABASE(), CONNECTION_ID(), VERSION()/*!50038 , @@hostname*/')),
'Connection info:', $dbh->{mysql_hostinfo},
'Character set info:', Dumper($dbh->selectall_arrayref(
"SHOW VARIABLES LIKE 'character_set%'", { Slice => {}})),
'$DBD::mysql::VERSION:', $DBD::mysql::VERSION,
'$DBI::VERSION:', $DBI::VERSION,
);
return $dbh;
}
sub get_hostname {
my ( $self, $dbh ) = @_;
if ( my ($host) = ($dbh->{mysql_hostinfo} || '') =~ m/^(\w+) via/ ) {
return $host;
}
my ( $hostname, $one ) = $dbh->selectrow_array(
'SELECT /*!50038 @@hostname, */ 1');
return $hostname;
}
sub disconnect {
my ( $self, $dbh ) = @_;
PTDEBUG && $self->print_active_handles($dbh);
$dbh->disconnect;
}
sub print_active_handles {
my ( $self, $thing, $level ) = @_;
$level ||= 0;
printf("# Active %sh: %s %s %s\n", ($thing->{Type} || 'undef'), "\t" x $level,
$thing, (($thing->{Type} || '') eq 'st' ? $thing->{Statement} || '' : ''))
or die "Cannot print: $OS_ERROR";
foreach my $handle ( grep {defined} @{ $thing->{ChildHandles} } ) {
$self->print_active_handles( $handle, $level + 1 );
}
}
sub copy {
my ( $self, $dsn_1, $dsn_2, %args ) = @_;
die 'I need a dsn_1 argument' unless $dsn_1;
die 'I need a dsn_2 argument' unless $dsn_2;
my %new_dsn = map {
my $key = $_;
my $val;
if ( $args{overwrite} ) {
$val = defined $dsn_1->{$key} ? $dsn_1->{$key} : $dsn_2->{$key};
}
else {
$val = defined $dsn_2->{$key} ? $dsn_2->{$key} : $dsn_1->{$key};
}
$key => $val;
} keys %{$self->{opts}};
return \%new_dsn;
}
sub set_vars {
my ($self, $dbh, $vars) = @_;
return unless $vars;
foreach my $var ( sort keys %$vars ) {
my $val = $vars->{$var}->{val};
(my $quoted_var = $var) =~ s/_/\\_/;
my ($var_exists, $current_val);
eval {
($var_exists, $current_val) = $dbh->selectrow_array(
"SHOW VARIABLES LIKE '$quoted_var'");
};
my $e = $EVAL_ERROR;
if ( $e ) {
PTDEBUG && _d($e);
}
if ( $vars->{$var}->{default} && !$var_exists ) {
PTDEBUG && _d('Not setting default var', $var,
'because it does not exist');
next;
}
if ( $current_val && $current_val eq $val ) {
PTDEBUG && _d('Not setting var', $var, 'because its value',
'is already', $val);
next;
}
my $sql = "SET SESSION $var=$val";
PTDEBUG && _d($dbh, $sql);
eval { $dbh->do($sql) };
if ( my $set_error = $EVAL_ERROR ) {
chomp($set_error);
$set_error =~ s/ at \S+ line \d+//;
my $msg = "Error setting $var: $set_error";
if ( $current_val ) {
$msg .= " The current value for $var is $current_val. "
. "If the variable is read only (not dynamic), specify "
. "--set-vars $var=$current_val to avoid this warning, "
. "else manually set the variable and restart MySQL.";
}
warn $msg . "\n\n";
}
}
return;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End DSNParser package
# ###########################################################################
# ###########################################################################
# OptionParser package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/OptionParser.pm
# t/lib/OptionParser.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package OptionParser;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use List::Util qw(max);
use Getopt::Long;
use Data::Dumper;
my $POD_link_re = '[LC]<"?([^">]+)"?>';
sub new {
my ( $class, %args ) = @_;
my @required_args = qw();
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
}
my ($program_name) = $PROGRAM_NAME =~ m/([.A-Za-z-]+)$/;
$program_name ||= $PROGRAM_NAME;
my $home = $ENV{HOME} || $ENV{HOMEPATH} || $ENV{USERPROFILE} || '.';
my %attributes = (
'type' => 1,
'short form' => 1,
'group' => 1,
'default' => 1,
'cumulative' => 1,
'negatable' => 1,
);
my $self = {
head1 => 'OPTIONS', # These args are used internally
skip_rules => 0, # to instantiate another Option-
item => '--(.*)', # Parser obj that parses the
attributes => \%attributes, # DSN OPTIONS section. Tools
parse_attributes => \&_parse_attribs, # don't tinker with these args.
%args,
strict => 1, # disabled by a special rule
program_name => $program_name,
opts => {},
got_opts => 0,
short_opts => {},
defaults => {},
groups => {},
allowed_groups => {},
errors => [],
rules => [], # desc of rules for --help
mutex => [], # rule: opts are mutually exclusive
atleast1 => [], # rule: at least one opt is required
disables => {}, # rule: opt disables other opts
defaults_to => {}, # rule: opt defaults to value of other opt
DSNParser => undef,
default_files => [
"/etc/percona-toolkit/percona-toolkit.conf",
"/etc/percona-toolkit/$program_name.conf",
"$home/.percona-toolkit.conf",
"$home/.$program_name.conf",
],
types => {
string => 's', # standard Getopt type
int => 'i', # standard Getopt type
float => 'f', # standard Getopt type
Hash => 'H', # hash, formed from a comma-separated list
hash => 'h', # hash as above, but only if a value is given
Array => 'A', # array, similar to Hash
array => 'a', # array, similar to hash
DSN => 'd', # DSN
size => 'z', # size with kMG suffix (powers of 2^10)
time => 'm', # time, with an optional suffix of s/h/m/d
},
};
return bless $self, $class;
}
sub get_specs {
my ( $self, $file ) = @_;
$file ||= $self->{file} || __FILE__;
my @specs = $self->_pod_to_specs($file);
$self->_parse_specs(@specs);
open my $fh, "<", $file or die "Cannot open $file: $OS_ERROR";
my $contents = do { local $/ = undef; <$fh> };
close $fh;
if ( $contents =~ m/^=head1 DSN OPTIONS/m ) {
PTDEBUG && _d('Parsing DSN OPTIONS');
my $dsn_attribs = {
dsn => 1,
copy => 1,
};
my $parse_dsn_attribs = sub {
my ( $self, $option, $attribs ) = @_;
map {
my $val = $attribs->{$_};
if ( $val ) {
$val = $val eq 'yes' ? 1
: $val eq 'no' ? 0
: $val;
$attribs->{$_} = $val;
}
} keys %$attribs;
return {
key => $option,
%$attribs,
};
};
my $dsn_o = new OptionParser(
description => 'DSN OPTIONS',
head1 => 'DSN OPTIONS',
dsn => 0, # XXX don't infinitely recurse!
item => '\* (.)', # key opts are a single character
skip_rules => 1, # no rules before opts
attributes => $dsn_attribs,
parse_attributes => $parse_dsn_attribs,
);
my @dsn_opts = map {
my $opts = {
key => $_->{spec}->{key},
dsn => $_->{spec}->{dsn},
copy => $_->{spec}->{copy},
desc => $_->{desc},
};
$opts;
} $dsn_o->_pod_to_specs($file);
$self->{DSNParser} = DSNParser->new(opts => \@dsn_opts);
}
if ( $contents =~ m/^=head1 VERSION\n\n^(.+)$/m ) {
$self->{version} = $1;
PTDEBUG && _d($self->{version});
}
return;
}
sub DSNParser {
my ( $self ) = @_;
return $self->{DSNParser};
};
sub get_defaults_files {
my ( $self ) = @_;
return @{$self->{default_files}};
}
sub _pod_to_specs {
my ( $self, $file ) = @_;
$file ||= $self->{file} || __FILE__;
open my $fh, '<', $file or die "Cannot open $file: $OS_ERROR";
my @specs = ();
my @rules = ();
my $para;
local $INPUT_RECORD_SEPARATOR = '';
while ( $para = <$fh> ) {
next unless $para =~ m/^=head1 $self->{head1}/;
last;
}
while ( $para = <$fh> ) {
last if $para =~ m/^=over/;
next if $self->{skip_rules};
chomp $para;
$para =~ s/\s+/ /g;
$para =~ s/$POD_link_re/$1/go;
PTDEBUG && _d('Option rule:', $para);
push @rules, $para;
}
die "POD has no $self->{head1} section" unless $para;
do {
if ( my ($option) = $para =~ m/^=item $self->{item}/ ) {
chomp $para;
PTDEBUG && _d($para);
my %attribs;
$para = <$fh>; # read next paragraph, possibly attributes
if ( $para =~ m/: / ) { # attributes
$para =~ s/\s+\Z//g;
%attribs = map {
my ( $attrib, $val) = split(/: /, $_);
die "Unrecognized attribute for --$option: $attrib"
unless $self->{attributes}->{$attrib};
($attrib, $val);
} split(/; /, $para);
if ( $attribs{'short form'} ) {
$attribs{'short form'} =~ s/-//;
}
$para = <$fh>; # read next paragraph, probably short help desc
}
else {
PTDEBUG && _d('Option has no attributes');
}
$para =~ s/\s+\Z//g;
$para =~ s/\s+/ /g;
$para =~ s/$POD_link_re/$1/go;
$para =~ s/\.(?:\n.*| [A-Z].*|\Z)//s;
PTDEBUG && _d('Short help:', $para);
die "No description after option spec $option" if $para =~ m/^=item/;
if ( my ($base_option) = $option =~ m/^\[no\](.*)/ ) {
$option = $base_option;
$attribs{'negatable'} = 1;
}
push @specs, {
spec => $self->{parse_attributes}->($self, $option, \%attribs),
desc => $para
. (defined $attribs{default} ? " (default $attribs{default})" : ''),
group => ($attribs{'group'} ? $attribs{'group'} : 'default'),
};
}
while ( $para = <$fh> ) {
last unless $para;
if ( $para =~ m/^=head1/ ) {
$para = undef; # Can't 'last' out of a do {} block.
last;
}
last if $para =~ m/^=item /;
}
} while ( $para );
die "No valid specs in $self->{head1}" unless @specs;
close $fh;
return @specs, @rules;
}
sub _parse_specs {
my ( $self, @specs ) = @_;
my %disables; # special rule that requires deferred checking
foreach my $opt ( @specs ) {
if ( ref $opt ) { # It's an option spec, not a rule.
PTDEBUG && _d('Parsing opt spec:',
map { ($_, '=>', $opt->{$_}) } keys %$opt);
my ( $long, $short ) = $opt->{spec} =~ m/^([\w-]+)(?:\|([^!+=]*))?/;
if ( !$long ) {
die "Cannot parse long option from spec $opt->{spec}";
}
$opt->{long} = $long;
die "Duplicate long option --$long" if exists $self->{opts}->{$long};
$self->{opts}->{$long} = $opt;
if ( length $long == 1 ) {
PTDEBUG && _d('Long opt', $long, 'looks like short opt');
$self->{short_opts}->{$long} = $long;
}
if ( $short ) {
die "Duplicate short option -$short"
if exists $self->{short_opts}->{$short};
$self->{short_opts}->{$short} = $long;
$opt->{short} = $short;
}
else {
$opt->{short} = undef;
}
$opt->{is_negatable} = $opt->{spec} =~ m/!/ ? 1 : 0;
$opt->{is_cumulative} = $opt->{spec} =~ m/\+/ ? 1 : 0;
$opt->{is_required} = $opt->{desc} =~ m/required/ ? 1 : 0;
$opt->{group} ||= 'default';
$self->{groups}->{ $opt->{group} }->{$long} = 1;
$opt->{value} = undef;
$opt->{got} = 0;
my ( $type ) = $opt->{spec} =~ m/=(.)/;
$opt->{type} = $type;
PTDEBUG && _d($long, 'type:', $type);
$opt->{spec} =~ s/=./=s/ if ( $type && $type =~ m/[HhAadzm]/ );
if ( (my ($def) = $opt->{desc} =~ m/default\b(?: ([^)]+))?/) ) {
$self->{defaults}->{$long} = defined $def ? $def : 1;
PTDEBUG && _d($long, 'default:', $def);
}
if ( $long eq 'config' ) {
$self->{defaults}->{$long} = join(',', $self->get_defaults_files());
}
if ( (my ($dis) = $opt->{desc} =~ m/(disables .*)/) ) {
$disables{$long} = $dis;
PTDEBUG && _d('Deferring check of disables rule for', $opt, $dis);
}
$self->{opts}->{$long} = $opt;
}
else { # It's an option rule, not a spec.
PTDEBUG && _d('Parsing rule:', $opt);
push @{$self->{rules}}, $opt;
my @participants = $self->_get_participants($opt);
my $rule_ok = 0;
if ( $opt =~ m/mutually exclusive|one and only one/ ) {
$rule_ok = 1;
push @{$self->{mutex}}, \@participants;
PTDEBUG && _d(@participants, 'are mutually exclusive');
}
if ( $opt =~ m/at least one|one and only one/ ) {
$rule_ok = 1;
push @{$self->{atleast1}}, \@participants;
PTDEBUG && _d(@participants, 'require at least one');
}
if ( $opt =~ m/default to/ ) {
$rule_ok = 1;
$self->{defaults_to}->{$participants[0]} = $participants[1];
PTDEBUG && _d($participants[0], 'defaults to', $participants[1]);
}
if ( $opt =~ m/restricted to option groups/ ) {
$rule_ok = 1;
my ($groups) = $opt =~ m/groups ([\w\s\,]+)/;
my @groups = split(',', $groups);
%{$self->{allowed_groups}->{$participants[0]}} = map {
s/\s+//;
$_ => 1;
} @groups;
}
if( $opt =~ m/accepts additional command-line arguments/ ) {
$rule_ok = 1;
$self->{strict} = 0;
PTDEBUG && _d("Strict mode disabled by rule");
}
die "Unrecognized option rule: $opt" unless $rule_ok;
}
}
foreach my $long ( keys %disables ) {
my @participants = $self->_get_participants($disables{$long});
$self->{disables}->{$long} = \@participants;
PTDEBUG && _d('Option', $long, 'disables', @participants);
}
return;
}
sub _get_participants {
my ( $self, $str ) = @_;
my @participants;
foreach my $long ( $str =~ m/--(?:\[no\])?([\w-]+)/g ) {
die "Option --$long does not exist while processing rule $str"
unless exists $self->{opts}->{$long};
push @participants, $long;
}
PTDEBUG && _d('Participants for', $str, ':', @participants);
return @participants;
}
sub opts {
my ( $self ) = @_;
my %opts = %{$self->{opts}};
return %opts;
}
sub short_opts {
my ( $self ) = @_;
my %short_opts = %{$self->{short_opts}};
return %short_opts;
}
sub set_defaults {
my ( $self, %defaults ) = @_;
$self->{defaults} = {};
foreach my $long ( keys %defaults ) {
die "Cannot set default for nonexistent option $long"
unless exists $self->{opts}->{$long};
$self->{defaults}->{$long} = $defaults{$long};
PTDEBUG && _d('Default val for', $long, ':', $defaults{$long});
}
return;
}
sub get_defaults {
my ( $self ) = @_;
return $self->{defaults};
}
sub get_groups {
my ( $self ) = @_;
return $self->{groups};
}
sub _set_option {
my ( $self, $opt, $val ) = @_;
my $long = exists $self->{opts}->{$opt} ? $opt
: exists $self->{short_opts}->{$opt} ? $self->{short_opts}->{$opt}
: die "Getopt::Long gave a nonexistent option: $opt";
$opt = $self->{opts}->{$long};
if ( $opt->{is_cumulative} ) {
$opt->{value}++;
}
elsif ( ($opt->{type} || '') eq 's' && $val =~ m/^--?(.+)/ ) {
my $next_opt = $1;
if ( exists $self->{opts}->{$next_opt}
|| exists $self->{short_opts}->{$next_opt} ) {
$self->save_error("--$long requires a string value");
return;
}
else {
$opt->{value} = $val;
}
}
else {
$opt->{value} = $val;
}
$opt->{got} = 1;
PTDEBUG && _d('Got option', $long, '=', $val);
}
sub get_opts {
my ( $self ) = @_;
foreach my $long ( keys %{$self->{opts}} ) {
$self->{opts}->{$long}->{got} = 0;
$self->{opts}->{$long}->{value}
= exists $self->{defaults}->{$long} ? $self->{defaults}->{$long}
: $self->{opts}->{$long}->{is_cumulative} ? 0
: undef;
}
$self->{got_opts} = 0;
$self->{errors} = [];
if ( @ARGV && $ARGV[0] eq "--config" ) {
shift @ARGV;
$self->_set_option('config', shift @ARGV);
}
if ( $self->has('config') ) {
my @extra_args;
foreach my $filename ( split(',', $self->get('config')) ) {
eval {
push @extra_args, $self->_read_config_file($filename);
};
if ( $EVAL_ERROR ) {
if ( $self->got('config') ) {
die $EVAL_ERROR;
}
elsif ( PTDEBUG ) {
_d($EVAL_ERROR);
}
}
}
unshift @ARGV, @extra_args;
}
Getopt::Long::Configure('no_ignore_case', 'bundling');
GetOptions(
map { $_->{spec} => sub { $self->_set_option(@_); } }
grep { $_->{long} ne 'config' } # --config is handled specially above.
values %{$self->{opts}}
) or $self->save_error('Error parsing options');
if ( exists $self->{opts}->{version} && $self->{opts}->{version}->{got} ) {
if ( $self->{version} ) {
print $self->{version}, "\n";
}
else {
print "Error parsing version. See the VERSION section of the tool's documentation.\n";
}
exit 1;
}
if ( @ARGV && $self->{strict} ) {
$self->save_error("Unrecognized command-line options @ARGV");
}
foreach my $mutex ( @{$self->{mutex}} ) {
my @set = grep { $self->{opts}->{$_}->{got} } @$mutex;
if ( @set > 1 ) {
my $err = join(', ', map { "--$self->{opts}->{$_}->{long}" }
@{$mutex}[ 0 .. scalar(@$mutex) - 2] )
. ' and --'.$self->{opts}->{$mutex->[-1]}->{long}
. ' are mutually exclusive.';
$self->save_error($err);
}
}
foreach my $required ( @{$self->{atleast1}} ) {
my @set = grep { $self->{opts}->{$_}->{got} } @$required;
if ( @set == 0 ) {
my $err = join(', ', map { "--$self->{opts}->{$_}->{long}" }
@{$required}[ 0 .. scalar(@$required) - 2] )
.' or --'.$self->{opts}->{$required->[-1]}->{long};
$self->save_error("Specify at least one of $err");
}
}
$self->_check_opts( keys %{$self->{opts}} );
$self->{got_opts} = 1;
return;
}
sub _check_opts {
my ( $self, @long ) = @_;
my $long_last = scalar @long;
while ( @long ) {
foreach my $i ( 0..$#long ) {
my $long = $long[$i];
next unless $long;
my $opt = $self->{opts}->{$long};
if ( $opt->{got} ) {
if ( exists $self->{disables}->{$long} ) {
my @disable_opts = @{$self->{disables}->{$long}};
map { $self->{opts}->{$_}->{value} = undef; } @disable_opts;
PTDEBUG && _d('Unset options', @disable_opts,
'because', $long,'disables them');
}
if ( exists $self->{allowed_groups}->{$long} ) {
my @restricted_groups = grep {
!exists $self->{allowed_groups}->{$long}->{$_}
} keys %{$self->{groups}};
my @restricted_opts;
foreach my $restricted_group ( @restricted_groups ) {
RESTRICTED_OPT:
foreach my $restricted_opt (
keys %{$self->{groups}->{$restricted_group}} )
{
next RESTRICTED_OPT if $restricted_opt eq $long;
push @restricted_opts, $restricted_opt
if $self->{opts}->{$restricted_opt}->{got};
}
}
if ( @restricted_opts ) {
my $err;
if ( @restricted_opts == 1 ) {
$err = "--$restricted_opts[0]";
}
else {
$err = join(', ',
map { "--$self->{opts}->{$_}->{long}" }
grep { $_ }
@restricted_opts[0..scalar(@restricted_opts) - 2]
)
. ' or --'.$self->{opts}->{$restricted_opts[-1]}->{long};
}
$self->save_error("--$long is not allowed with $err");
}
}
}
elsif ( $opt->{is_required} ) {
$self->save_error("Required option --$long must be specified");
}
$self->_validate_type($opt);
if ( $opt->{parsed} ) {
delete $long[$i];
}
else {
PTDEBUG && _d('Temporarily failed to parse', $long);
}
}
die "Failed to parse options, possibly due to circular dependencies"
if @long == $long_last;
$long_last = @long;
}
return;
}
sub _validate_type {
my ( $self, $opt ) = @_;
return unless $opt;
if ( !$opt->{type} ) {
$opt->{parsed} = 1;
return;
}
my $val = $opt->{value};
if ( $val && $opt->{type} eq 'm' ) { # type time
PTDEBUG && _d('Parsing option', $opt->{long}, 'as a time value');
my ( $prefix, $num, $suffix ) = $val =~ m/([+-]?)(\d+)([a-z])?$/;
if ( !$suffix ) {
my ( $s ) = $opt->{desc} =~ m/\(suffix (.)\)/;
$suffix = $s || 's';
PTDEBUG && _d('No suffix given; using', $suffix, 'for',
$opt->{long}, '(value:', $val, ')');
}
if ( $suffix =~ m/[smhd]/ ) {
$val = $suffix eq 's' ? $num # Seconds
: $suffix eq 'm' ? $num * 60 # Minutes
: $suffix eq 'h' ? $num * 3600 # Hours
: $num * 86400; # Days
$opt->{value} = ($prefix || '') . $val;
PTDEBUG && _d('Setting option', $opt->{long}, 'to', $val);
}
else {
$self->save_error("Invalid time suffix for --$opt->{long}");
}
}
elsif ( $val && $opt->{type} eq 'd' ) { # type DSN
PTDEBUG && _d('Parsing option', $opt->{long}, 'as a DSN');
my $prev = {};
my $from_key = $self->{defaults_to}->{ $opt->{long} };
if ( $from_key ) {
PTDEBUG && _d($opt->{long}, 'DSN copies from', $from_key, 'DSN');
if ( $self->{opts}->{$from_key}->{parsed} ) {
$prev = $self->{opts}->{$from_key}->{value};
}
else {
PTDEBUG && _d('Cannot parse', $opt->{long}, 'until',
$from_key, 'parsed');
return;
}
}
my $defaults = $self->{DSNParser}->parse_options($self);
$opt->{value} = $self->{DSNParser}->parse($val, $prev, $defaults);
}
elsif ( $val && $opt->{type} eq 'z' ) { # type size
PTDEBUG && _d('Parsing option', $opt->{long}, 'as a size value');
$self->_parse_size($opt, $val);
}
elsif ( $opt->{type} eq 'H' || (defined $val && $opt->{type} eq 'h') ) {
$opt->{value} = { map { $_ => 1 } split(/(?<!\\),\s*/, ($val || '')) };
}
elsif ( $opt->{type} eq 'A' || (defined $val && $opt->{type} eq 'a') ) {
$opt->{value} = [ split(/(?<!\\),\s*/, ($val || '')) ];
}
else {
PTDEBUG && _d('Nothing to validate for option',
$opt->{long}, 'type', $opt->{type}, 'value', $val);
}
$opt->{parsed} = 1;
return;
}
sub get {
my ( $self, $opt ) = @_;
my $long = (length $opt == 1 ? $self->{short_opts}->{$opt} : $opt);
die "Option $opt does not exist"
unless $long && exists $self->{opts}->{$long};
return $self->{opts}->{$long}->{value};
}
sub got {
my ( $self, $opt ) = @_;
my $long = (length $opt == 1 ? $self->{short_opts}->{$opt} : $opt);
die "Option $opt does not exist"
unless $long && exists $self->{opts}->{$long};
return $self->{opts}->{$long}->{got};
}
sub has {
my ( $self, $opt ) = @_;
my $long = (length $opt == 1 ? $self->{short_opts}->{$opt} : $opt);
return defined $long ? exists $self->{opts}->{$long} : 0;
}
sub set {
my ( $self, $opt, $val ) = @_;
my $long = (length $opt == 1 ? $self->{short_opts}->{$opt} : $opt);
die "Option $opt does not exist"
unless $long && exists $self->{opts}->{$long};
$self->{opts}->{$long}->{value} = $val;
return;
}
sub save_error {
my ( $self, $error ) = @_;
push @{$self->{errors}}, $error;
return;
}
sub errors {
my ( $self ) = @_;
return $self->{errors};
}
sub usage {
my ( $self ) = @_;
warn "No usage string is set" unless $self->{usage}; # XXX
return "Usage: " . ($self->{usage} || '') . "\n";
}
sub descr {
my ( $self ) = @_;
warn "No description string is set" unless $self->{description}; # XXX
my $descr = ($self->{description} || $self->{program_name} || '')
. " For more details, please use the --help option, "
. "or try 'perldoc $PROGRAM_NAME' "
. "for complete documentation.";
$descr = join("\n", $descr =~ m/(.{0,80})(?:\s+|$)/g)
unless $ENV{DONT_BREAK_LINES};
$descr =~ s/ +$//mg;
return $descr;
}
sub usage_or_errors {
my ( $self, $file, $return ) = @_;
$file ||= $self->{file} || __FILE__;
if ( !$self->{description} || !$self->{usage} ) {
PTDEBUG && _d("Getting description and usage from SYNOPSIS in", $file);
my %synop = $self->_parse_synopsis($file);
$self->{description} ||= $synop{description};
$self->{usage} ||= $synop{usage};
PTDEBUG && _d("Description:", $self->{description},
"\nUsage:", $self->{usage});
}
if ( $self->{opts}->{help}->{got} ) {
print $self->print_usage() or die "Cannot print usage: $OS_ERROR";
exit 0 unless $return;
}
elsif ( scalar @{$self->{errors}} ) {
print $self->print_errors() or die "Cannot print errors: $OS_ERROR";
exit 1 unless $return;
}
return;
}
sub print_errors {
my ( $self ) = @_;
my $usage = $self->usage() . "\n";
if ( (my @errors = @{$self->{errors}}) ) {
$usage .= join("\n * ", 'Errors in command-line arguments:', @errors)
. "\n";
}
return $usage . "\n" . $self->descr();
}
sub print_usage {
my ( $self ) = @_;
die "Run get_opts() before print_usage()" unless $self->{got_opts};
my @opts = values %{$self->{opts}};
my $maxl = max(
map {
length($_->{long}) # option long name
+ ($_->{is_negatable} ? 4 : 0) # "[no]" if opt is negatable
+ ($_->{type} ? 2 : 0) # "=x" where x is the opt type
}
@opts);
my $maxs = max(0,
map {
length($_)
+ ($self->{opts}->{$_}->{is_negatable} ? 4 : 0)
+ ($self->{opts}->{$_}->{type} ? 2 : 0)
}
values %{$self->{short_opts}});
my $lcol = max($maxl, ($maxs + 3));
my $rcol = 80 - $lcol - 6;
my $rpad = ' ' x ( 80 - $rcol );
$maxs = max($lcol - 3, $maxs);
my $usage = $self->descr() . "\n" . $self->usage();
my @groups = reverse sort grep { $_ ne 'default'; } keys %{$self->{groups}};
push @groups, 'default';
foreach my $group ( reverse @groups ) {
$usage .= "\n".($group eq 'default' ? 'Options' : $group).":\n\n";
foreach my $opt (
sort { $a->{long} cmp $b->{long} }
grep { $_->{group} eq $group }
@opts )
{
my $long = $opt->{is_negatable} ? "[no]$opt->{long}" : $opt->{long};
my $short = $opt->{short};
my $desc = $opt->{desc};
$long .= $opt->{type} ? "=$opt->{type}" : "";
if ( $opt->{type} && $opt->{type} eq 'm' ) {
my ($s) = $desc =~ m/\(suffix (.)\)/;
$s ||= 's';
$desc =~ s/\s+\(suffix .\)//;
$desc .= ". Optional suffix s=seconds, m=minutes, h=hours, "
. "d=days; if no suffix, $s is used.";
}
$desc = join("\n$rpad", grep { $_ } $desc =~ m/(.{0,$rcol}(?!\W))(?:\s+|(?<=\W)|$)/g);
$desc =~ s/ +$//mg;
if ( $short ) {
$usage .= sprintf(" --%-${maxs}s -%s %s\n", $long, $short, $desc);
}
else {
$usage .= sprintf(" --%-${lcol}s %s\n", $long, $desc);
}
}
}
$usage .= "\nOption types: s=string, i=integer, f=float, h/H/a/A=comma-separated list, d=DSN, z=size, m=time\n";
if ( (my @rules = @{$self->{rules}}) ) {
$usage .= "\nRules:\n\n";
$usage .= join("\n", map { " $_" } @rules) . "\n";
}
if ( $self->{DSNParser} ) {
$usage .= "\n" . $self->{DSNParser}->usage();
}
$usage .= "\nOptions and values after processing arguments:\n\n";
foreach my $opt ( sort { $a->{long} cmp $b->{long} } @opts ) {
my $val = $opt->{value};
my $type = $opt->{type} || '';
my $bool = $opt->{spec} =~ m/^[\w-]+(?:\|[\w-])?!?$/;
$val = $bool ? ( $val ? 'TRUE' : 'FALSE' )
: !defined $val ? '(No value)'
: $type eq 'd' ? $self->{DSNParser}->as_string($val)
: $type =~ m/H|h/ ? join(',', sort keys %$val)
: $type =~ m/A|a/ ? join(',', @$val)
: $val;
$usage .= sprintf(" --%-${lcol}s %s\n", $opt->{long}, $val);
}
return $usage;
}
sub prompt_noecho {
shift @_ if ref $_[0] eq __PACKAGE__;
my ( $prompt ) = @_;
local $OUTPUT_AUTOFLUSH = 1;
print STDERR $prompt
or die "Cannot print: $OS_ERROR";
my $response;
eval {
require Term::ReadKey;
Term::ReadKey::ReadMode('noecho');
chomp($response = <STDIN>);
Term::ReadKey::ReadMode('normal');
print "\n"
or die "Cannot print: $OS_ERROR";
};
if ( $EVAL_ERROR ) {
die "Cannot read response; is Term::ReadKey installed? $EVAL_ERROR";
}
return $response;
}
sub _read_config_file {
my ( $self, $filename ) = @_;
open my $fh, "<", $filename or die "Cannot open $filename: $OS_ERROR\n";
my @args;
my $prefix = '--';
my $parse = 1;
LINE:
while ( my $line = <$fh> ) {
chomp $line;
next LINE if $line =~ m/^\s*(?:\#|\;|$)/;
$line =~ s/\s+#.*$//g;
$line =~ s/^\s+|\s+$//g;
if ( $line eq '--' ) {
$prefix = '';
$parse = 0;
next LINE;
}
if ( $parse
&& (my($opt, $arg) = $line =~ m/^\s*([^=\s]+?)(?:\s*=\s*(.*?)\s*)?$/)
) {
push @args, grep { defined $_ } ("$prefix$opt", $arg);
}
elsif ( $line =~ m/./ ) {
push @args, $line;
}
else {
die "Syntax error in file $filename at line $INPUT_LINE_NUMBER";
}
}
close $fh;
return @args;
}
sub read_para_after {
my ( $self, $file, $regex ) = @_;
open my $fh, "<", $file or die "Can't open $file: $OS_ERROR";
local $INPUT_RECORD_SEPARATOR = '';
my $para;
while ( $para = <$fh> ) {
next unless $para =~ m/^=pod$/m;
last;
}
while ( $para = <$fh> ) {
next unless $para =~ m/$regex/;
last;
}
$para = <$fh>;
chomp($para);
close $fh or die "Can't close $file: $OS_ERROR";
return $para;
}
sub clone {
my ( $self ) = @_;
my %clone = map {
my $hashref = $self->{$_};
my $val_copy = {};
foreach my $key ( keys %$hashref ) {
my $ref = ref $hashref->{$key};
$val_copy->{$key} = !$ref ? $hashref->{$key}
: $ref eq 'HASH' ? { %{$hashref->{$key}} }
: $ref eq 'ARRAY' ? [ @{$hashref->{$key}} ]
: $hashref->{$key};
}
$_ => $val_copy;
} qw(opts short_opts defaults);
foreach my $scalar ( qw(got_opts) ) {
$clone{$scalar} = $self->{$scalar};
}
return bless \%clone;
}
sub _parse_size {
my ( $self, $opt, $val ) = @_;
if ( lc($val || '') eq 'null' ) {
PTDEBUG && _d('NULL size for', $opt->{long});
$opt->{value} = 'null';
return;
}
my %factor_for = (k => 1_024, M => 1_048_576, G => 1_073_741_824);
my ($pre, $num, $factor) = $val =~ m/^([+-])?(\d+)([kMG])?$/;
if ( defined $num ) {
if ( $factor ) {
$num *= $factor_for{$factor};
PTDEBUG && _d('Setting option', $opt->{y},
'to num', $num, '* factor', $factor);
}
$opt->{value} = ($pre || '') . $num;
}
else {
$self->save_error("Invalid size for --$opt->{long}: $val");
}
return;
}
sub _parse_attribs {
my ( $self, $option, $attribs ) = @_;
my $types = $self->{types};
return $option
. ($attribs->{'short form'} ? '|' . $attribs->{'short form'} : '' )
. ($attribs->{'negatable'} ? '!' : '' )
. ($attribs->{'cumulative'} ? '+' : '' )
. ($attribs->{'type'} ? '=' . $types->{$attribs->{type}} : '' );
}
sub _parse_synopsis {
my ( $self, $file ) = @_;
$file ||= $self->{file} || __FILE__;
PTDEBUG && _d("Parsing SYNOPSIS in", $file);
local $INPUT_RECORD_SEPARATOR = ''; # read paragraphs
open my $fh, "<", $file or die "Cannot open $file: $OS_ERROR";
my $para;
1 while defined($para = <$fh>) && $para !~ m/^=head1 SYNOPSIS/;
die "$file does not contain a SYNOPSIS section" unless $para;
my @synop;
for ( 1..2 ) { # 1 for the usage, 2 for the description
my $para = <$fh>;
push @synop, $para;
}
close $fh;
PTDEBUG && _d("Raw SYNOPSIS text:", @synop);
my ($usage, $desc) = @synop;
die "The SYNOPSIS section in $file is not formatted properly"
unless $usage && $desc;
$usage =~ s/^\s*Usage:\s+(.+)/$1/;
chomp $usage;
$desc =~ s/\n/ /g;
$desc =~ s/\s{2,}/ /g;
$desc =~ s/\. ([A-Z][a-z])/. $1/g;
$desc =~ s/\s+$//;
return (
description => $desc,
usage => $usage,
);
};
sub set_vars {
my ($self, $file) = @_;
$file ||= $self->{file} || __FILE__;
my %user_vars;
my $user_vars = $self->has('set-vars') ? $self->get('set-vars') : undef;
if ( $user_vars ) {
foreach my $var_val ( @$user_vars ) {
my ($var, $val) = $var_val =~ m/([^\s=]+)=(\S+)/;
die "Invalid --set-vars value: $var_val\n" unless $var && defined $val;
$user_vars{$var} = {
val => $val,
default => 0,
};
}
}
my %default_vars;
my $default_vars = $self->read_para_after($file, qr/MAGIC_set_vars/);
if ( $default_vars ) {
%default_vars = map {
my $var_val = $_;
my ($var, $val) = $var_val =~ m/([^\s=]+)=(\S+)/;
die "Invalid --set-vars value: $var_val\n" unless $var && defined $val;
$var => {
val => $val,
default => 1,
};
} split("\n", $default_vars);
}
my %vars = (
%default_vars, # first the tool's defaults
%user_vars, # then the user's which overwrite the defaults
);
PTDEBUG && _d('--set-vars:', Dumper(\%vars));
return \%vars;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
if ( PTDEBUG ) {
print STDERR '# ', $^X, ' ', $], "\n";
if ( my $uname = `uname -a` ) {
$uname =~ s/\s+/ /g;
print STDERR "# $uname\n";
}
print STDERR '# Arguments: ',
join(' ', map { my $a = "_[$_]_"; $a =~ s/\n/\n# /g; $a; } @ARGV), "\n";
}
1;
}
# ###########################################################################
# End OptionParser package
# ###########################################################################
# ###########################################################################
# KeySize package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/KeySize.pm
# t/lib/KeySize.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package KeySize;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
sub new {
my ( $class, %args ) = @_;
my $self = { %args };
return bless $self, $class;
}
sub get_key_size {
my ( $self, %args ) = @_;
foreach my $arg ( qw(name cols tbl_name tbl_struct dbh) ) {
die "I need a $arg argument" unless $args{$arg};
}
my $name = $args{name};
my @cols = @{$args{cols}};
my $dbh = $args{dbh};
$self->{explain} = '';
$self->{query} = '';
$self->{error} = '';
if ( @cols == 0 ) {
$self->{error} = "No columns for key $name";
return;
}
my $key_exists = $self->_key_exists(%args);
PTDEBUG && _d('Key', $name, 'exists in', $args{tbl_name}, ':',
$key_exists ? 'yes': 'no');
my $sql = 'EXPLAIN SELECT ' . join(', ', @cols)
. ' FROM ' . $args{tbl_name}
. ($key_exists ? " FORCE INDEX (`$name`)" : '')
. ' WHERE ';
my @where_cols;
foreach my $col ( @cols ) {
push @where_cols, "$col=1";
}
if ( scalar(@cols) == 1 && !$args{only_eq} ) {
push @where_cols, "$cols[0]<>1";
}
$sql .= join(' OR ', @where_cols);
$self->{query} = $sql;
PTDEBUG && _d('sql:', $sql);
my $explain;
my $sth = $dbh->prepare($sql);
eval { $sth->execute(); };
if ( $EVAL_ERROR ) {
chomp $EVAL_ERROR;
$self->{error} = "Cannot get size of $name key: $EVAL_ERROR";
return;
}
$explain = $sth->fetchrow_hashref();
$self->{explain} = $explain;
my $key_len = $explain->{key_len};
my $rows = $explain->{rows};
my $chosen_key = $explain->{key}; # May differ from $name
PTDEBUG && _d('MySQL chose key:', $chosen_key, 'len:', $key_len,
'rows:', $rows);
if ( $chosen_key && $key_len eq '0' ) {
if ( $args{recurse} ) {
$self->{error} = "key_len = 0 in EXPLAIN:\n"
. _explain_to_text($explain);
return;
}
else {
return $self->get_key_size(
%args,
only_eq => 1,
recurse => 1,
);
}
}
my $key_size = 0;
if ( $key_len && $rows ) {
if ( $chosen_key =~ m/,/ && $key_len =~ m/,/ ) {
$self->{error} = "MySQL chose multiple keys: $chosen_key";
return;
}
$key_size = $key_len * $rows;
}
else {
$self->{error} = "key_len or rows NULL in EXPLAIN:\n"
. _explain_to_text($explain);
return;
}
return $key_size, $chosen_key;
}
sub query {
my ( $self ) = @_;
return $self->{query};
}
sub explain {
my ( $self ) = @_;
return _explain_to_text($self->{explain});
}
sub error {
my ( $self ) = @_;
return $self->{error};
}
sub _key_exists {
my ( $self, %args ) = @_;
return exists $args{tbl_struct}->{keys}->{ lc $args{name} } ? 1 : 0;
}
sub _explain_to_text {
my ( $explain ) = @_;
return join("\n",
map { "$_: ".($explain->{$_} ? $explain->{$_} : 'NULL') }
sort keys %$explain
);
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End KeySize package
# ###########################################################################
# ###########################################################################
# DuplicateKeyFinder package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/DuplicateKeyFinder.pm
# t/lib/DuplicateKeyFinder.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package DuplicateKeyFinder;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
sub new {
my ( $class, %args ) = @_;
my $self = {};
return bless $self, $class;
}
sub get_duplicate_keys {
my ( $self, $keys, %args ) = @_;
die "I need a keys argument" unless $keys;
my %keys = %$keys; # Copy keys because we remove non-duplicates.
my $primary_key;
my @unique_keys;
my @normal_keys;
my @fulltext_keys;
my @dupes;
KEY:
# sort by key name for consistent testing
foreach my $key ( sort {$a->{name} cmp $b->{name}} values %keys ) {
$key->{real_cols} = [ @{$key->{cols}} ];
$key->{len_cols} = length $key->{colnames};
if ( $key->{name} eq 'PRIMARY'
|| ($args{clustered_key} && $key->{name} eq $args{clustered_key}) ) {
$primary_key = $key;
PTDEBUG && _d('primary key:', $key->{name});
next KEY;
}
my $is_fulltext = $key->{type} eq 'FULLTEXT' ? 1 : 0;
if ( $args{ignore_order} || $is_fulltext ) {
my $ordered_cols = join(',', sort(split(/,/, $key->{colnames})));
PTDEBUG && _d('Reordered', $key->{name}, 'cols from',
$key->{colnames}, 'to', $ordered_cols);
$key->{colnames} = $ordered_cols;
}
my $push_to = $key->{is_unique} ? \@unique_keys : \@normal_keys;
if ( !$args{ignore_structure} ) {
$push_to = \@fulltext_keys if $is_fulltext;
}
push @$push_to, $key;
}
push @normal_keys, $self->unconstrain_keys($primary_key, \@unique_keys);
if ( $primary_key ) {
PTDEBUG && _d('Comparing PRIMARY KEY to UNIQUE keys');
push @dupes,
$self->remove_prefix_duplicates([$primary_key], \@unique_keys, %args);
PTDEBUG && _d('Comparing PRIMARY KEY to normal keys');
push @dupes,
$self->remove_prefix_duplicates([$primary_key], \@normal_keys, %args);
}
PTDEBUG && _d('Comparing UNIQUE keys to normal keys');
push @dupes,
$self->remove_prefix_duplicates(\@unique_keys, \@normal_keys, %args);
PTDEBUG && _d('Comparing normal keys');
push @dupes,
$self->remove_prefix_duplicates(\@normal_keys, \@normal_keys, %args);
PTDEBUG && _d('Comparing FULLTEXT keys');
push @dupes,
$self->remove_prefix_duplicates(\@fulltext_keys, \@fulltext_keys, %args, exact_duplicates => 1);
my $clustered_key = $args{clustered_key} ? $keys{$args{clustered_key}}
: undef;
PTDEBUG && _d('clustered key:',
$clustered_key ? ($clustered_key->{name}, $clustered_key->{colnames})
: 'none');
if ( $clustered_key
&& $args{clustered}
&& $args{tbl_info}->{engine}
&& $args{tbl_info}->{engine} =~ m/InnoDB/i )
{
PTDEBUG && _d('Removing UNIQUE dupes of clustered key');
push @dupes,
$self->remove_clustered_duplicates($clustered_key, \@unique_keys, %args);
PTDEBUG && _d('Removing ordinary dupes of clustered key');
push @dupes,
$self->remove_clustered_duplicates($clustered_key, \@normal_keys, %args);
}
return \@dupes;
}
sub get_duplicate_fks {
my ( $self, $fks, %args ) = @_;
die "I need a fks argument" unless $fks;
# sort by name for consistent testing
my @fks = sort {$a->{name} cmp $b->{name}} values %$fks;
my @dupes;
foreach my $i ( 0..$#fks - 1 ) {
next unless $fks[$i];
foreach my $j ( $i+1..$#fks ) {
next unless $fks[$j];
my $i_cols = join(',', sort @{$fks[$i]->{cols}} );
my $j_cols = join(',', sort @{$fks[$j]->{cols}} );
my $i_pcols = join(',', sort @{$fks[$i]->{parent_cols}} );
my $j_pcols = join(',', sort @{$fks[$j]->{parent_cols}} );
if ( $fks[$i]->{parent_tblname} eq $fks[$j]->{parent_tblname}
&& $i_cols eq $j_cols
&& $i_pcols eq $j_pcols ) {
my $dupe = {
key => $fks[$j]->{name},
cols => [ @{$fks[$j]->{cols}} ],
ddl => $fks[$j]->{ddl},
duplicate_of => $fks[$i]->{name},
duplicate_of_cols => [ @{$fks[$i]->{cols}} ],
duplicate_of_ddl => $fks[$i]->{ddl},
reason =>
"FOREIGN KEY $fks[$j]->{name} ($fks[$j]->{colnames}) "
. "REFERENCES $fks[$j]->{parent_tblname} "
. "($fks[$j]->{parent_colnames}) "
. 'is a duplicate of '
. "FOREIGN KEY $fks[$i]->{name} ($fks[$i]->{colnames}) "
. "REFERENCES $fks[$i]->{parent_tblname} "
."($fks[$i]->{parent_colnames})",
dupe_type => 'fk',
};
push @dupes, $dupe;
delete $fks[$j];
$args{callback}->($dupe, %args) if $args{callback};
}
}
}
return \@dupes;
}
sub remove_prefix_duplicates {
my ( $self, $left_keys, $right_keys, %args ) = @_;
my @dupes;
my $right_offset;
my $last_left_key;
my $last_right_key = scalar(@$right_keys) - 1;
if ( $right_keys != $left_keys ) {
@$left_keys = sort { lc($a->{colnames}) cmp lc($b->{colnames}) }
grep { defined $_; }
@$left_keys;
@$right_keys = sort { lc($a->{colnames}) cmp lc($b->{colnames}) }
grep { defined $_; }
@$right_keys;
$last_left_key = scalar(@$left_keys) - 1;
$right_offset = 0;
}
else {
@$left_keys = reverse sort { lc($a->{colnames}) cmp lc($b->{colnames}) }
grep { defined $_; }
@$left_keys;
$last_left_key = scalar(@$left_keys) - 2;
$right_offset = 1;
}
LEFT_KEY:
foreach my $left_index ( 0..$last_left_key ) {
next LEFT_KEY unless defined $left_keys->[$left_index];
RIGHT_KEY:
foreach my $right_index ( $left_index+$right_offset..$last_right_key ) {
next RIGHT_KEY unless defined $right_keys->[$right_index];
my $left_name = $left_keys->[$left_index]->{name};
my $left_cols = $left_keys->[$left_index]->{colnames};
my $left_len_cols = $left_keys->[$left_index]->{len_cols};
my $right_name = $right_keys->[$right_index]->{name};
my $right_cols = $right_keys->[$right_index]->{colnames};
my $right_len_cols = $right_keys->[$right_index]->{len_cols};
PTDEBUG && _d('Comparing left', $left_name, '(',$left_cols,')',
'to right', $right_name, '(',$right_cols,')');
if ( substr($left_cols, 0, $right_len_cols)
eq substr($right_cols, 0, $right_len_cols) ) {
if ( $args{exact_duplicates} && ($right_len_cols<$left_len_cols) ) {
PTDEBUG && _d($right_name, 'not exact duplicate of', $left_name);
next RIGHT_KEY;
}
if ( exists $right_keys->[$right_index]->{unique_col} ) {
PTDEBUG && _d('Cannot remove', $right_name,
'because is constrains col',
$right_keys->[$right_index]->{cols}->[0]);
next RIGHT_KEY;
}
PTDEBUG && _d('Remove', $right_name);
my $reason;
if ( my $type = $right_keys->[$right_index]->{unconstrained} ) {
$reason .= "Uniqueness of $right_name ignored because "
. $right_keys->[$right_index]->{constraining_key}->{name}
. " is a $type constraint\n";
}
my $exact_dupe = $right_len_cols < $left_len_cols ? 0 : 1;
$reason .= $right_name
. ($exact_dupe ? ' is a duplicate of '
: ' is a left-prefix of ')
. $left_name;
my $dupe = {
key => $right_name,
cols => $right_keys->[$right_index]->{real_cols},
ddl => $right_keys->[$right_index]->{ddl},
duplicate_of => $left_name,
duplicate_of_cols => $left_keys->[$left_index]->{real_cols},
duplicate_of_ddl => $left_keys->[$left_index]->{ddl},
reason => $reason,
dupe_type => $exact_dupe ? 'exact' : 'prefix',
};
push @dupes, $dupe;
delete $right_keys->[$right_index];
$args{callback}->($dupe, %args) if $args{callback};
}
else {
PTDEBUG && _d($right_name, 'not left-prefix of', $left_name);
next RIGHT_KEY;
}
} # RIGHT_KEY
} # LEFT_KEY
PTDEBUG && _d('No more keys');
@$left_keys = grep { defined $_; } @$left_keys;
@$right_keys = grep { defined $_; } @$right_keys;
return @dupes;
}
sub remove_clustered_duplicates {
my ( $self, $ck, $keys, %args ) = @_;
die "I need a ck argument" unless $ck;
die "I need a keys argument" unless $keys;
my $ck_cols = $ck->{colnames};
my @dupes;
KEY:
for my $i ( 0 .. @$keys - 1 ) {
my $key = $keys->[$i]->{colnames};
if ( $key =~ m/$ck_cols$/ ) {
PTDEBUG && _d("clustered key dupe:", $keys->[$i]->{name},
$keys->[$i]->{colnames});
my $dupe = {
key => $keys->[$i]->{name},
cols => $keys->[$i]->{real_cols},
ddl => $keys->[$i]->{ddl},
duplicate_of => $ck->{name},
duplicate_of_cols => $ck->{real_cols},
duplicate_of_ddl => $ck->{ddl},
reason => "Key $keys->[$i]->{name} ends with a "
. "prefix of the clustered index",
dupe_type => 'clustered',
short_key => $self->shorten_clustered_duplicate(
$ck_cols,
join(',', map { "`$_`" }
@{$keys->[$i]->{real_cols}})
),
};
push @dupes, $dupe;
delete $keys->[$i];
$args{callback}->($dupe, %args) if $args{callback};
}
}
PTDEBUG && _d('No more keys');
@$keys = grep { defined $_; } @$keys;
return @dupes;
}
sub shorten_clustered_duplicate {
my ( $self, $ck_cols, $dupe_key_cols ) = @_;
return $ck_cols if $ck_cols eq $dupe_key_cols;
$dupe_key_cols =~ s/$ck_cols$//;
$dupe_key_cols =~ s/,+$//;
return $dupe_key_cols;
}
sub unconstrain_keys {
my ( $self, $primary_key, $unique_keys ) = @_;
die "I need a unique_keys argument" unless $unique_keys;
my %unique_cols;
my @unique_sets;
my %unconstrain;
my @unconstrained_keys;
PTDEBUG && _d('Unconstraining redundantly unique keys');
UNIQUE_KEY:
foreach my $unique_key ( $primary_key, @$unique_keys ) {
next unless $unique_key; # primary key may be undefined
my $cols = $unique_key->{cols};
if ( @$cols == 1 ) {
if ( !exists $unique_cols{$cols->[0]} ) {
PTDEBUG && _d($unique_key->{name}, 'defines unique column:',
$cols->[0]);
$unique_cols{$cols->[0]} = $unique_key;
$unique_key->{unique_col} = 1;
}
else {
PTDEBUG && _d($unique_key->{name},
'redundantly constrains unique column:', $cols->[0]);
$unique_key->{exact_dupe} = 1;
$unique_key->{constraining_key} = $unique_cols{$cols->[0]};
}
}
else {
local $LIST_SEPARATOR = '-';
PTDEBUG && _d($unique_key->{name}, 'defines unique set:', @$cols);
push @unique_sets, { cols => $cols, key => $unique_key };
}
}
UNIQUE_SET:
foreach my $unique_set ( @unique_sets ) {
my $n_unique_cols = 0;
COL:
foreach my $col ( @{$unique_set->{cols}} ) {
if ( exists $unique_cols{$col} ) {
PTDEBUG && _d('Unique set', $unique_set->{key}->{name},
'has unique col', $col);
last COL if ++$n_unique_cols > 1;
$unique_set->{constraining_key} = $unique_cols{$col};
}
}
if ( $n_unique_cols && $unique_set->{key}->{name} ne 'PRIMARY' ) {
PTDEBUG && _d('Will unconstrain unique set',
$unique_set->{key}->{name},
'because it is redundantly constrained by key',
$unique_set->{constraining_key}->{name},
'(',$unique_set->{constraining_key}->{colnames},')');
$unconstrain{$unique_set->{key}->{name}}
= $unique_set->{constraining_key};
}
}
for my $i ( 0..(scalar @$unique_keys-1) ) {
if ( exists $unconstrain{$unique_keys->[$i]->{name}} ) {
PTDEBUG && _d('Unconstraining weak', $unique_keys->[$i]->{name});
$unique_keys->[$i]->{unconstrained} = 'stronger';
$unique_keys->[$i]->{constraining_key}
= $unconstrain{$unique_keys->[$i]->{name}};
push @unconstrained_keys, $unique_keys->[$i];
delete $unique_keys->[$i];
}
elsif ( $unique_keys->[$i]->{exact_dupe} ) {
PTDEBUG && _d('Unconstraining dupe', $unique_keys->[$i]->{name});
$unique_keys->[$i]->{unconstrained} = 'duplicate';
push @unconstrained_keys, $unique_keys->[$i];
delete $unique_keys->[$i];
}
}
PTDEBUG && _d('No more keys');
return @unconstrained_keys;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End DuplicateKeyFinder package
# ###########################################################################
# ###########################################################################
# Daemon package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/Daemon.pm
# t/lib/Daemon.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package Daemon;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use POSIX qw(setsid);
use Fcntl qw(:DEFAULT);
sub new {
my ($class, %args) = @_;
my $self = {
log_file => $args{log_file},
pid_file => $args{pid_file},
daemonize => $args{daemonize},
force_log_file => $args{force_log_file},
parent_exit => $args{parent_exit},
pid_file_owner => 0,
};
return bless $self, $class;
}
sub run {
my ($self) = @_;
my $daemonize = $self->{daemonize};
my $pid_file = $self->{pid_file};
my $log_file = $self->{log_file};
my $force_log_file = $self->{force_log_file};
my $parent_exit = $self->{parent_exit};
PTDEBUG && _d('Starting daemon');
if ( $pid_file ) {
eval {
$self->_make_pid_file(
pid => $PID, # parent's pid
pid_file => $pid_file,
);
};
die "$EVAL_ERROR\n" if $EVAL_ERROR;
if ( !$daemonize ) {
$self->{pid_file_owner} = $PID; # parent's pid
}
}
if ( $daemonize ) {
defined (my $child_pid = fork()) or die "Cannot fork: $OS_ERROR";
if ( $child_pid ) {
PTDEBUG && _d('Forked child', $child_pid);
$parent_exit->($child_pid) if $parent_exit;
exit 0;
}
POSIX::setsid() or die "Cannot start a new session: $OS_ERROR";
chdir '/' or die "Cannot chdir to /: $OS_ERROR";
if ( $pid_file ) {
$self->_update_pid_file(
pid => $PID, # child's pid
pid_file => $pid_file,
);
$self->{pid_file_owner} = $PID;
}
}
if ( $daemonize || $force_log_file ) {
PTDEBUG && _d('Redirecting STDIN to /dev/null');
close STDIN;
open STDIN, '/dev/null'
or die "Cannot reopen STDIN to /dev/null: $OS_ERROR";
if ( $log_file ) {
PTDEBUG && _d('Redirecting STDOUT and STDERR to', $log_file);
close STDOUT;
open STDOUT, '>>', $log_file
or die "Cannot open log file $log_file: $OS_ERROR";
close STDERR;
open STDERR, ">&STDOUT"
or die "Cannot dupe STDERR to STDOUT: $OS_ERROR";
}
else {
if ( -t STDOUT ) {
PTDEBUG && _d('No log file and STDOUT is a terminal;',
'redirecting to /dev/null');
close STDOUT;
open STDOUT, '>', '/dev/null'
or die "Cannot reopen STDOUT to /dev/null: $OS_ERROR";
}
if ( -t STDERR ) {
PTDEBUG && _d('No log file and STDERR is a terminal;',
'redirecting to /dev/null');
close STDERR;
open STDERR, '>', '/dev/null'
or die "Cannot reopen STDERR to /dev/null: $OS_ERROR";
}
}
$OUTPUT_AUTOFLUSH = 1;
}
PTDEBUG && _d('Daemon running');
return;
}
sub _make_pid_file {
my ($self, %args) = @_;
my @required_args = qw(pid pid_file);
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
};
my $pid = $args{pid};
my $pid_file = $args{pid_file};
eval {
sysopen(PID_FH, $pid_file, O_RDWR|O_CREAT|O_EXCL) or die $OS_ERROR;
print PID_FH $PID, "\n";
close PID_FH;
};
if ( my $e = $EVAL_ERROR ) {
if ( $e =~ m/file exists/i ) {
my $old_pid = $self->_check_pid_file(
pid_file => $pid_file,
pid => $PID,
);
if ( $old_pid ) {
warn "Overwriting PID file $pid_file because PID $old_pid "
. "is not running.\n";
}
$self->_update_pid_file(
pid => $PID,
pid_file => $pid_file
);
}
else {
die "Error creating PID file $pid_file: $e\n";
}
}
return;
}
sub _check_pid_file {
my ($self, %args) = @_;
my @required_args = qw(pid_file pid);
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
};
my $pid_file = $args{pid_file};
my $pid = $args{pid};
PTDEBUG && _d('Checking if PID in', $pid_file, 'is running');
if ( ! -f $pid_file ) {
PTDEBUG && _d('PID file', $pid_file, 'does not exist');
return;
}
open my $fh, '<', $pid_file
or die "Error opening $pid_file: $OS_ERROR";
my $existing_pid = do { local $/; <$fh> };
chomp($existing_pid) if $existing_pid;
close $fh
or die "Error closing $pid_file: $OS_ERROR";
if ( $existing_pid ) {
if ( $existing_pid == $pid ) {
warn "The current PID $pid already holds the PID file $pid_file\n";
return;
}
else {
PTDEBUG && _d('Checking if PID', $existing_pid, 'is running');
my $pid_is_alive = kill 0, $existing_pid;
if ( $pid_is_alive ) {
die "PID file $pid_file exists and PID $existing_pid is running\n";
}
}
}
else {
die "PID file $pid_file exists but it is empty. Remove the file "
. "if the process is no longer running.\n";
}
return $existing_pid;
}
sub _update_pid_file {
my ($self, %args) = @_;
my @required_args = qw(pid pid_file);
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
};
my $pid = $args{pid};
my $pid_file = $args{pid_file};
open my $fh, '>', $pid_file
or die "Cannot open $pid_file: $OS_ERROR";
print { $fh } $pid, "\n"
or die "Cannot print to $pid_file: $OS_ERROR";
close $fh
or warn "Cannot close $pid_file: $OS_ERROR";
return;
}
sub remove_pid_file {
my ($self, $pid_file) = @_;
$pid_file ||= $self->{pid_file};
if ( $pid_file && -f $pid_file ) {
unlink $self->{pid_file}
or warn "Cannot remove PID file $pid_file: $OS_ERROR";
PTDEBUG && _d('Removed PID file');
}
else {
PTDEBUG && _d('No PID to remove');
}
return;
}
sub DESTROY {
my ($self) = @_;
if ( $self->{pid_file_owner} == $PID ) {
$self->remove_pid_file();
}
return;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End Daemon package
# ###########################################################################
# ###########################################################################
# Schema package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/Schema.pm
# t/lib/Schema.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package Schema;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
sub new {
my ( $class, %args ) = @_;
my @required_args = qw();
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
}
my $self = {
%args,
schema => {}, # keyed on db->tbl
};
return bless $self, $class;
}
sub get_schema {
my ( $self ) = @_;
return $self->{schema};
}
sub get_table {
my ( $self, $db_name, $tbl_name ) = @_;
if ( exists $self->{schema}->{$db_name}
&& exists $self->{schema}->{$db_name}->{$tbl_name} ) {
return $self->{schema}->{$db_name}->{$tbl_name};
}
return;
}
sub add_schema_object {
my ( $self, $schema_object ) = @_;
die "I need a schema_object argument" unless $schema_object;
my ($db, $tbl) = @{$schema_object}{qw(db tbl)};
if ( !$db || !$tbl ) {
warn "No database or table for schema object";
return;
}
my $tbl_struct = $schema_object->{tbl_struct};
if ( !$tbl_struct ) {
warn "No table structure for $db.$tbl";
return;
}
$self->{schema}->{lc $db}->{lc $tbl} = $schema_object;
return;
}
sub find_column {
my ( $self, %args ) = @_;
my $ignore = $args{ignore};
my $schema = $self->{schema};
my ($col, $tbl, $db);
if ( my $col_name = $args{col_name} ) {
($col, $tbl, $db) = reverse map { s/`//g; $_ } split /[.]/, $col_name;
PTDEBUG && _d('Column', $col_name, 'has db', $db, 'tbl', $tbl,
'col', $col);
}
else {
($col, $tbl, $db) = @args{qw(col tbl db)};
}
$db = lc($db || '');
$tbl = lc($tbl || '');
$col = lc($col || '');
if ( !$col ) {
PTDEBUG && _d('No column specified or parsed');
return;
}
PTDEBUG && _d('Finding column', $col, 'in', $db, $tbl);
if ( $db && !$schema->{$db} ) {
PTDEBUG && _d('Database', $db, 'does not exist');
return;
}
if ( $db && $tbl && !$schema->{$db}->{$tbl} ) {
PTDEBUG && _d('Table', $tbl, 'does not exist in database', $db);
return;
}
my @tbls;
my @search_dbs = $db ? ($db) : keys %$schema;
DATABASE:
foreach my $search_db ( @search_dbs ) {
my @search_tbls = $tbl ? ($tbl) : keys %{$schema->{$search_db}};
TABLE:
foreach my $search_tbl ( @search_tbls ) {
next DATABASE unless exists $schema->{$search_db}->{$search_tbl};
if ( $ignore
&& grep { $_->{db} eq $search_db && $_->{tbl} eq $search_tbl }
@$ignore ) {
PTDEBUG && _d('Ignoring', $search_db, $search_tbl, $col);
next TABLE;
}
my $tbl = $schema->{$search_db}->{$search_tbl};
if ( $tbl->{tbl_struct}->{is_col}->{$col} ) {
PTDEBUG && _d('Column', $col, 'exists in', $tbl->{db}, $tbl->{tbl});
push @tbls, $tbl;
}
}
}
return \@tbls;
}
sub find_table {
my ( $self, %args ) = @_;
my $ignore = $args{ignore};
my $schema = $self->{schema};
my ($tbl, $db);
if ( my $tbl_name = $args{tbl_name} ) {
($tbl, $db) = reverse map { s/`//g; $_ } split /[.]/, $tbl_name;
PTDEBUG && _d('Table', $tbl_name, 'has db', $db, 'tbl', $tbl);
}
else {
($tbl, $db) = @args{qw(tbl db)};
}
$db = lc($db || '');
$tbl = lc($tbl || '');
if ( !$tbl ) {
PTDEBUG && _d('No table specified or parsed');
return;
}
PTDEBUG && _d('Finding table', $tbl, 'in', $db);
if ( $db && !$schema->{$db} ) {
PTDEBUG && _d('Database', $db, 'does not exist');
return;
}
if ( $db && $tbl && !$schema->{$db}->{$tbl} ) {
PTDEBUG && _d('Table', $tbl, 'does not exist in database', $db);
return;
}
my @dbs;
my @search_dbs = $db ? ($db) : keys %$schema;
DATABASE:
foreach my $search_db ( @search_dbs ) {
if ( $ignore && grep { $_->{db} eq $search_db } @$ignore ) {
PTDEBUG && _d('Ignoring', $search_db);
next DATABASE;
}
if ( exists $schema->{$search_db}->{$tbl} ) {
PTDEBUG && _d('Table', $tbl, 'exists in', $search_db);
push @dbs, $search_db;
}
}
return \@dbs;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End Schema package
# ###########################################################################
# ###########################################################################
# SchemaIterator package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/SchemaIterator.pm
# t/lib/SchemaIterator.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package SchemaIterator;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use Data::Dumper;
$Data::Dumper::Indent = 1;
$Data::Dumper::Sortkeys = 1;
$Data::Dumper::Quotekeys = 0;
my $open_comment = qr{/\*!\d{5} };
my $tbl_name = qr{
CREATE\s+
(?:TEMPORARY\s+)?
TABLE\s+
(?:IF NOT EXISTS\s+)?
([^\(]+)
}x;
sub new {
my ( $class, %args ) = @_;
my @required_args = qw(OptionParser TableParser Quoter);
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
}
my ($file_itr, $dbh) = @args{qw(file_itr dbh)};
die "I need either a dbh or file_itr argument"
if (!$dbh && !$file_itr) || ($dbh && $file_itr);
my %resume;
if ( my $table = $args{resume} ) {
PTDEBUG && _d('Will resume from or after', $table);
my ($db, $tbl) = $args{Quoter}->split_unquote($table);
die "Resume table must be database-qualified: $table"
unless $db && $tbl;
$resume{db} = $db;
$resume{tbl} = $tbl;
}
my $self = {
%args,
resume => \%resume,
filters => _make_filters(%args),
};
return bless $self, $class;
}
sub _make_filters {
my ( %args ) = @_;
my @required_args = qw(OptionParser Quoter);
foreach my $arg ( @required_args ) {
die "I need a $arg argument" unless $args{$arg};
}
my ($o, $q) = @args{@required_args};
my %filters;
my @simple_filters = qw(
databases tables engines
ignore-databases ignore-tables ignore-engines);
FILTER:
foreach my $filter ( @simple_filters ) {
if ( $o->has($filter) ) {
my $objs = $o->get($filter);
next FILTER unless $objs && scalar keys %$objs;
my $is_table = $filter =~ m/table/ ? 1 : 0;
foreach my $obj ( keys %$objs ) {
die "Undefined value for --$filter" unless $obj;
$obj = lc $obj;
if ( $is_table ) {
my ($db, $tbl) = $q->split_unquote($obj);
$db ||= '*';
PTDEBUG && _d('Filter', $filter, 'value:', $db, $tbl);
$filters{$filter}->{$db}->{$tbl} = 1;
}
else { # database
PTDEBUG && _d('Filter', $filter, 'value:', $obj);
$filters{$filter}->{$obj} = 1;
}
}
}
}
my @regex_filters = qw(
databases-regex tables-regex
ignore-databases-regex ignore-tables-regex);
REGEX_FILTER:
foreach my $filter ( @regex_filters ) {
if ( $o->has($filter) ) {
my $pat = $o->get($filter);
next REGEX_FILTER unless $pat;
$filters{$filter} = qr/$pat/;
PTDEBUG && _d('Filter', $filter, 'value:', $filters{$filter});
}
}
PTDEBUG && _d('Schema object filters:', Dumper(\%filters));
return \%filters;
}
sub next {
my ( $self ) = @_;
if ( !$self->{initialized} ) {
$self->{initialized} = 1;
if ( $self->{resume}->{tbl} ) {
if ( !$self->table_is_allowed(@{$self->{resume}}{qw(db tbl)}) ) {
PTDEBUG && _d('Will resume after',
join('.', @{$self->{resume}}{qw(db tbl)}));
$self->{resume}->{after}->{tbl} = 1;
}
if ( !$self->database_is_allowed($self->{resume}->{db}) ) {
PTDEBUG && _d('Will resume after', $self->{resume}->{db});
$self->{resume}->{after}->{db} = 1;
}
}
}
my $schema_obj;
if ( $self->{file_itr} ) {
$schema_obj= $self->_iterate_files();
}
else { # dbh
$schema_obj= $self->_iterate_dbh();
}
if ( $schema_obj ) {
if ( my $schema = $self->{Schema} ) {
$schema->add_schema_object($schema_obj);
}
PTDEBUG && _d('Next schema object:',
$schema_obj->{db}, $schema_obj->{tbl});
}
return $schema_obj;
}
sub _iterate_files {
my ( $self ) = @_;
if ( !$self->{fh} ) {
my ($fh, $file) = $self->{file_itr}->();
if ( !$fh ) {
PTDEBUG && _d('No more files to iterate');
return;
}
$self->{fh} = $fh;
$self->{file} = $file;
}
my $fh = $self->{fh};
PTDEBUG && _d('Getting next schema object from', $self->{file});
local $INPUT_RECORD_SEPARATOR = '';
CHUNK:
while (defined(my $chunk = <$fh>)) {
if ($chunk =~ m/Database: (\S+)/) {
my $db = $1; # XXX
$db =~ s/^`//; # strip leading `
$db =~ s/`$//; # and trailing `
if ( $self->database_is_allowed($db)
&& $self->_resume_from_database($db) ) {
$self->{db} = $db;
}
}
elsif ($self->{db} && $chunk =~ m/CREATE TABLE/) {
if ($chunk =~ m/DROP VIEW IF EXISTS/) {
PTDEBUG && _d('Table is a VIEW, skipping');
next CHUNK;
}
my ($tbl) = $chunk =~ m/$tbl_name/;
$tbl =~ s/^\s*`//;
$tbl =~ s/`\s*$//;
if ( $self->_resume_from_table($tbl)
&& $self->table_is_allowed($self->{db}, $tbl) ) {
my ($ddl) = $chunk =~ m/^(?:$open_comment)?(CREATE TABLE.+?;)$/ms;
if ( !$ddl ) {
warn "Failed to parse CREATE TABLE from\n" . $chunk;
next CHUNK;
}
$ddl =~ s/ \*\/;\Z/;/; # remove end of version comment
my $tbl_struct = $self->{TableParser}->parse($ddl);
if ( $self->engine_is_allowed($tbl_struct->{engine}) ) {
return {
db => $self->{db},
tbl => $tbl,
name => $self->{Quoter}->quote($self->{db}, $tbl),
ddl => $ddl,
tbl_struct => $tbl_struct,
};
}
}
}
} # CHUNK
PTDEBUG && _d('No more schema objects in', $self->{file});
close $self->{fh};
$self->{fh} = undef;
return $self->_iterate_files();
}
sub _iterate_dbh {
my ( $self ) = @_;
my $q = $self->{Quoter};
my $tp = $self->{TableParser};
my $dbh = $self->{dbh};
PTDEBUG && _d('Getting next schema object from dbh', $dbh);
if ( !defined $self->{dbs} ) {
my $sql = 'SHOW DATABASES';
PTDEBUG && _d($sql);
my @dbs = grep {
$self->_resume_from_database($_)
&&
$self->database_is_allowed($_)
} @{$dbh->selectcol_arrayref($sql)};
PTDEBUG && _d('Found', scalar @dbs, 'databases');
$self->{dbs} = \@dbs;
}
DATABASE:
while ( $self->{db} || defined(my $db = shift @{$self->{dbs}}) ) {
if ( !$self->{db} ) {
PTDEBUG && _d('Next database:', $db);
$self->{db} = $db;
}
if ( !$self->{tbls} ) {
my $sql = 'SHOW /*!50002 FULL*/ TABLES FROM ' . $q->quote($self->{db});
PTDEBUG && _d($sql);
my @tbls = map {
$_->[0]; # (tbl, type)
}
grep {
my ($tbl, $type) = @$_;
(!$type || ($type ne 'VIEW'))
&& $self->_resume_from_table($tbl)
&& $self->table_is_allowed($self->{db}, $tbl);
}
@{$dbh->selectall_arrayref($sql)};
PTDEBUG && _d('Found', scalar @tbls, 'tables in database',$self->{db});
$self->{tbls} = \@tbls;
}
TABLE:
while ( my $tbl = shift @{$self->{tbls}} ) {
my $ddl = eval { $tp->get_create_table($dbh, $self->{db}, $tbl) };
if ( my $e = $EVAL_ERROR ) {
my $table_name = "$self->{db}.$tbl";
if ( $e =~ /\QTable '$table_name' doesn't exist/ ) {
PTDEBUG && _d("$table_name no longer exists");
}
else {
warn "Skipping $table_name because SHOW CREATE TABLE failed: $e";
}
next TABLE;
}
my $tbl_struct = $tp->parse($ddl);
if ( $self->engine_is_allowed($tbl_struct->{engine}) ) {
return {
db => $self->{db},
tbl => $tbl,
name => $q->quote($self->{db}, $tbl),
ddl => $ddl,
tbl_struct => $tbl_struct,
};
}
}
PTDEBUG && _d('No more tables in database', $self->{db});
$self->{db} = undef;
$self->{tbls} = undef;
} # DATABASE
PTDEBUG && _d('No more databases');
return;
}
sub database_is_allowed {
my ( $self, $db ) = @_;
die "I need a db argument" unless $db;
$db = lc $db;
my $filter = $self->{filters};
if ( $db =~ m/information_schema|performance_schema|lost\+found/ ) {
PTDEBUG && _d('Database', $db, 'is a system database, ignoring');
return 0;
}
if ( $self->{filters}->{'ignore-databases'}->{$db} ) {
PTDEBUG && _d('Database', $db, 'is in --ignore-databases list');
return 0;
}
if ( $filter->{'ignore-databases-regex'}
&& $db =~ $filter->{'ignore-databases-regex'} ) {
PTDEBUG && _d('Database', $db, 'matches --ignore-databases-regex');
return 0;
}
if ( $filter->{'databases'}
&& !$filter->{'databases'}->{$db} ) {
PTDEBUG && _d('Database', $db, 'is not in --databases list, ignoring');
return 0;
}
if ( $filter->{'databases-regex'}
&& $db !~ $filter->{'databases-regex'} ) {
PTDEBUG && _d('Database', $db, 'does not match --databases-regex, ignoring');
return 0;
}
return 1;
}
sub table_is_allowed {
my ( $self, $db, $tbl ) = @_;
die "I need a db argument" unless $db;
die "I need a tbl argument" unless $tbl;
$db = lc $db;
$tbl = lc $tbl;
my $filter = $self->{filters};
return 0 if $db eq 'mysql' && $tbl =~ m/^(?:
general_log
|slow_log
|innodb_index_stats
|innodb_table_stats
|slave_master_info
|slave_relay_log_info
|slave_worker_info
)$/x;
if ( $filter->{'ignore-tables'}->{'*'}->{$tbl}
|| $filter->{'ignore-tables'}->{$db}->{$tbl}) {
PTDEBUG && _d('Table', $tbl, 'is in --ignore-tables list');
return 0;
}
if ( $filter->{'ignore-tables-regex'}
&& $tbl =~ $filter->{'ignore-tables-regex'} ) {
PTDEBUG && _d('Table', $tbl, 'matches --ignore-tables-regex');
return 0;
}
if ( $filter->{'tables'}
&& (!$filter->{'tables'}->{'*'}->{$tbl} && !$filter->{'tables'}->{$db}->{$tbl}) ) {
PTDEBUG && _d('Table', $tbl, 'is not in --tables list, ignoring');
return 0;
}
if ( $filter->{'tables-regex'}
&& $tbl !~ $filter->{'tables-regex'} ) {
PTDEBUG && _d('Table', $tbl, 'does not match --tables-regex, ignoring');
return 0;
}
if ( $filter->{'tables'}
&& $filter->{'tables'}->{$tbl}
&& $filter->{'tables'}->{$tbl} ne '*'
&& $filter->{'tables'}->{$tbl} ne $db ) {
PTDEBUG && _d('Table', $tbl, 'is only allowed in database',
$filter->{'tables'}->{$tbl});
return 0;
}
return 1;
}
sub engine_is_allowed {
my ( $self, $engine ) = @_;
if ( !$engine ) {
PTDEBUG && _d('No engine specified; allowing the table');
return 1;
}
$engine = lc $engine;
my $filter = $self->{filters};
if ( $filter->{'ignore-engines'}->{$engine} ) {
PTDEBUG && _d('Engine', $engine, 'is in --ignore-databases list');
return 0;
}
if ( $filter->{'engines'}
&& !$filter->{'engines'}->{$engine} ) {
PTDEBUG && _d('Engine', $engine, 'is not in --engines list, ignoring');
return 0;
}
return 1;
}
sub _resume_from_database {
my ($self, $db) = @_;
return 1 unless $self->{resume}->{db};
if ( $db eq $self->{resume}->{db} ) {
if ( !$self->{resume}->{after}->{db} ) {
PTDEBUG && _d('Resuming from db', $db);
delete $self->{resume}->{db};
return 1;
}
else {
PTDEBUG && _d('Resuming after db', $db);
delete $self->{resume}->{db};
delete $self->{resume}->{tbl};
}
}
return 0;
}
sub _resume_from_table {
my ($self, $tbl) = @_;
return 1 unless $self->{resume}->{tbl};
if ( $tbl eq $self->{resume}->{tbl} ) {
if ( !$self->{resume}->{after}->{tbl} ) {
PTDEBUG && _d('Resuming from table', $tbl);
delete $self->{resume}->{tbl};
return 1;
}
else {
PTDEBUG && _d('Resuming after table', $tbl);
delete $self->{resume}->{tbl};
}
}
return 0;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End SchemaIterator package
# ###########################################################################
# ###########################################################################
# HTTP::Micro package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/HTTP/Micro.pm
# t/lib/HTTP/Micro.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package HTTP::Micro;
our $VERSION = '0.01';
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use Carp ();
my @attributes;
BEGIN {
@attributes = qw(agent timeout);
no strict 'refs';
for my $accessor ( @attributes ) {
*{$accessor} = sub {
@_ > 1 ? $_[0]->{$accessor} = $_[1] : $_[0]->{$accessor};
};
}
}
sub new {
my($class, %args) = @_;
(my $agent = $class) =~ s{::}{-}g;
my $self = {
agent => $agent . "/" . ($class->VERSION || 0),
timeout => 60,
};
for my $key ( @attributes ) {
$self->{$key} = $args{$key} if exists $args{$key}
}
return bless $self, $class;
}
my %DefaultPort = (
http => 80,
https => 443,
);
sub request {
my ($self, $method, $url, $args) = @_;
@_ == 3 || (@_ == 4 && ref $args eq 'HASH')
or Carp::croak(q/Usage: $http->request(METHOD, URL, [HASHREF])/);
$args ||= {}; # we keep some state in this during _request
my $response;
for ( 0 .. 1 ) {
$response = eval { $self->_request($method, $url, $args) };
last unless $@ && $method eq 'GET'
&& $@ =~ m{^(?:Socket closed|Unexpected end)};
}
if (my $e = "$@") {
$response = {
success => q{},
status => 599,
reason => 'Internal Exception',
content => $e,
headers => {
'content-type' => 'text/plain',
'content-length' => length $e,
}
};
}
return $response;
}
sub _request {
my ($self, $method, $url, $args) = @_;
my ($scheme, $host, $port, $path_query) = $self->_split_url($url);
my $request = {
method => $method,
scheme => $scheme,
host_port => ($port == $DefaultPort{$scheme} ? $host : "$host:$port"),
uri => $path_query,
headers => {},
};
my $handle = HTTP::Micro::Handle->new(timeout => $self->{timeout});
$handle->connect($scheme, $host, $port);
$self->_prepare_headers_and_cb($request, $args);
$handle->write_request_header(@{$request}{qw/method uri headers/});
$handle->write_content_body($request) if $request->{content};
my $response;
do { $response = $handle->read_response_header }
until (substr($response->{status},0,1) ne '1');
if (!($method eq 'HEAD' || $response->{status} =~ /^[23]04/)) {
$response->{content} = '';
$handle->read_content_body(sub { $_[1]->{content} .= $_[0] }, $response);
}
$handle->close;
$response->{success} = substr($response->{status},0,1) eq '2';
return $response;
}
sub _prepare_headers_and_cb {
my ($self, $request, $args) = @_;
for ($args->{headers}) {
next unless defined;
while (my ($k, $v) = each %$_) {
$request->{headers}{lc $k} = $v;
}
}
$request->{headers}{'host'} = $request->{host_port};
$request->{headers}{'connection'} = "close";
$request->{headers}{'user-agent'} ||= $self->{agent};
if (defined $args->{content}) {
$request->{headers}{'content-type'} ||= "application/octet-stream";
utf8::downgrade($args->{content}, 1)
or Carp::croak(q/Wide character in request message body/);
$request->{headers}{'content-length'} = length $args->{content};
$request->{content} = $args->{content};
}
return;
}
sub _split_url {
my $url = pop;
my ($scheme, $authority, $path_query) = $url =~ m<\A([^:/?#]+)://([^/?#]*)([^#]*)>
or Carp::croak(qq/Cannot parse URL: '$url'/);
$scheme = lc $scheme;
$path_query = "/$path_query" unless $path_query =~ m<\A/>;
my $host = (length($authority)) ? lc $authority : 'localhost';
$host =~ s/\A[^@]*@//; # userinfo
my $port = do {
$host =~ s/:([0-9]*)\z// && length $1
? $1
: $DefaultPort{$scheme}
};
return ($scheme, $host, $port, $path_query);
}
} # HTTP::Micro
{
package HTTP::Micro::Handle;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use Carp qw(croak);
use Errno qw(EINTR EPIPE);
use IO::Socket qw(SOCK_STREAM);
sub BUFSIZE () { 32768 }
my $Printable = sub {
local $_ = shift;
s/\r/\\r/g;
s/\n/\\n/g;
s/\t/\\t/g;
s/([^\x20-\x7E])/sprintf('\\x%.2X', ord($1))/ge;
$_;
};
sub new {
my ($class, %args) = @_;
return bless {
rbuf => '',
timeout => 60,
max_line_size => 16384,
%args
}, $class;
}
my $ssl_verify_args = {
check_cn => "when_only",
wildcards_in_alt => "anywhere",
wildcards_in_cn => "anywhere"
};
sub connect {
@_ == 4 || croak(q/Usage: $handle->connect(scheme, host, port)/);
my ($self, $scheme, $host, $port) = @_;
if ( $scheme eq 'https' ) {
eval "require IO::Socket::SSL"
unless exists $INC{'IO/Socket/SSL.pm'};
croak(qq/IO::Socket::SSL must be installed for https support\n/)
unless $INC{'IO/Socket/SSL.pm'};
}
elsif ( $scheme ne 'http' ) {
croak(qq/Unsupported URL scheme '$scheme'\n/);
}
$self->{fh} = IO::Socket::INET->new(
PeerHost => $host,
PeerPort => $port,
Proto => 'tcp',
Type => SOCK_STREAM,
Timeout => $self->{timeout}
) or croak(qq/Could not connect to '$host:$port': $@/);
binmode($self->{fh})
or croak(qq/Could not binmode() socket: '$!'/);
if ( $scheme eq 'https') {
IO::Socket::SSL->start_SSL($self->{fh});
ref($self->{fh}) eq 'IO::Socket::SSL'
or die(qq/SSL connection failed for $host\n/);
if ( $self->{fh}->can("verify_hostname") ) {
$self->{fh}->verify_hostname( $host, $ssl_verify_args )
or die(qq/SSL certificate not valid for $host\n/);
}
else {
my $fh = $self->{fh};
_verify_hostname_of_cert($host, _peer_certificate($fh), $ssl_verify_args)
or die(qq/SSL certificate not valid for $host\n/);
}
}
$self->{host} = $host;
$self->{port} = $port;
return $self;
}
sub close {
@_ == 1 || croak(q/Usage: $handle->close()/);
my ($self) = @_;
CORE::close($self->{fh})
or croak(qq/Could not close socket: '$!'/);
}
sub write {
@_ == 2 || croak(q/Usage: $handle->write(buf)/);
my ($self, $buf) = @_;
my $len = length $buf;
my $off = 0;
local $SIG{PIPE} = 'IGNORE';
while () {
$self->can_write
or croak(q/Timed out while waiting for socket to become ready for writing/);
my $r = syswrite($self->{fh}, $buf, $len, $off);
if (defined $r) {
$len -= $r;
$off += $r;
last unless $len > 0;
}
elsif ($! == EPIPE) {
croak(qq/Socket closed by remote server: $!/);
}
elsif ($! != EINTR) {
croak(qq/Could not write to socket: '$!'/);
}
}
return $off;
}
sub read {
@_ == 2 || @_ == 3 || croak(q/Usage: $handle->read(len)/);
my ($self, $len) = @_;
my $buf = '';
my $got = length $self->{rbuf};
if ($got) {
my $take = ($got < $len) ? $got : $len;
$buf = substr($self->{rbuf}, 0, $take, '');
$len -= $take;
}
while ($len > 0) {
$self->can_read
or croak(q/Timed out while waiting for socket to become ready for reading/);
my $r = sysread($self->{fh}, $buf, $len, length $buf);
if (defined $r) {
last unless $r;
$len -= $r;
}
elsif ($! != EINTR) {
croak(qq/Could not read from socket: '$!'/);
}
}
if ($len) {
croak(q/Unexpected end of stream/);
}
return $buf;
}
sub readline {
@_ == 1 || croak(q/Usage: $handle->readline()/);
my ($self) = @_;
while () {
if ($self->{rbuf} =~ s/\A ([^\x0D\x0A]* \x0D?\x0A)//x) {
return $1;
}
$self->can_read
or croak(q/Timed out while waiting for socket to become ready for reading/);
my $r = sysread($self->{fh}, $self->{rbuf}, BUFSIZE, length $self->{rbuf});
if (defined $r) {
last unless $r;
}
elsif ($! != EINTR) {
croak(qq/Could not read from socket: '$!'/);
}
}
croak(q/Unexpected end of stream while looking for line/);
}
sub read_header_lines {
@_ == 1 || @_ == 2 || croak(q/Usage: $handle->read_header_lines([headers])/);
my ($self, $headers) = @_;
$headers ||= {};
my $lines = 0;
my $val;
while () {
my $line = $self->readline;
if ($line =~ /\A ([^\x00-\x1F\x7F:]+) : [\x09\x20]* ([^\x0D\x0A]*)/x) {
my ($field_name) = lc $1;
$val = \($headers->{$field_name} = $2);
}
elsif ($line =~ /\A [\x09\x20]+ ([^\x0D\x0A]*)/x) {
$val
or croak(q/Unexpected header continuation line/);
next unless length $1;
$$val .= ' ' if length $$val;
$$val .= $1;
}
elsif ($line =~ /\A \x0D?\x0A \z/x) {
last;
}
else {
croak(q/Malformed header line: / . $Printable->($line));
}
}
return $headers;
}
sub write_header_lines {
(@_ == 2 && ref $_[1] eq 'HASH') || croak(q/Usage: $handle->write_header_lines(headers)/);
my($self, $headers) = @_;
my $buf = '';
while (my ($k, $v) = each %$headers) {
my $field_name = lc $k;
$field_name =~ /\A [\x21\x23-\x27\x2A\x2B\x2D\x2E\x30-\x39\x41-\x5A\x5E-\x7A\x7C\x7E]+ \z/x
or croak(q/Invalid HTTP header field name: / . $Printable->($field_name));
$field_name =~ s/\b(\w)/\u$1/g;
$buf .= "$field_name: $v\x0D\x0A";
}
$buf .= "\x0D\x0A";
return $self->write($buf);
}
sub read_content_body {
@_ == 3 || @_ == 4 || croak(q/Usage: $handle->read_content_body(callback, response, [read_length])/);
my ($self, $cb, $response, $len) = @_;
$len ||= $response->{headers}{'content-length'};
croak("No content-length in the returned response, and this "
. "UA doesn't implement chunking") unless defined $len;
while ($len > 0) {
my $read = ($len > BUFSIZE) ? BUFSIZE : $len;
$cb->($self->read($read), $response);
$len -= $read;
}
return;
}
sub write_content_body {
@_ == 2 || croak(q/Usage: $handle->write_content_body(request)/);
my ($self, $request) = @_;
my ($len, $content_length) = (0, $request->{headers}{'content-length'});
$len += $self->write($request->{content});
$len == $content_length
or croak(qq/Content-Length missmatch (got: $len expected: $content_length)/);
return $len;
}
sub read_response_header {
@_ == 1 || croak(q/Usage: $handle->read_response_header()/);
my ($self) = @_;
my $line = $self->readline;
$line =~ /\A (HTTP\/(0*\d+\.0*\d+)) [\x09\x20]+ ([0-9]{3}) [\x09\x20]+ ([^\x0D\x0A]*) \x0D?\x0A/x
or croak(q/Malformed Status-Line: / . $Printable->($line));
my ($protocol, $version, $status, $reason) = ($1, $2, $3, $4);
return {
status => $status,
reason => $reason,
headers => $self->read_header_lines,
protocol => $protocol,
};
}
sub write_request_header {
@_ == 4 || croak(q/Usage: $handle->write_request_header(method, request_uri, headers)/);
my ($self, $method, $request_uri, $headers) = @_;
return $self->write("$method $request_uri HTTP/1.1\x0D\x0A")
+ $self->write_header_lines($headers);
}
sub _do_timeout {
my ($self, $type, $timeout) = @_;
$timeout = $self->{timeout}
unless defined $timeout && $timeout >= 0;
my $fd = fileno $self->{fh};
defined $fd && $fd >= 0
or croak(q/select(2): 'Bad file descriptor'/);
my $initial = time;
my $pending = $timeout;
my $nfound;
vec(my $fdset = '', $fd, 1) = 1;
while () {
$nfound = ($type eq 'read')
? select($fdset, undef, undef, $pending)
: select(undef, $fdset, undef, $pending) ;
if ($nfound == -1) {
$! == EINTR
or croak(qq/select(2): '$!'/);
redo if !$timeout || ($pending = $timeout - (time - $initial)) > 0;
$nfound = 0;
}
last;
}
$! = 0;
return $nfound;
}
sub can_read {
@_ == 1 || @_ == 2 || croak(q/Usage: $handle->can_read([timeout])/);
my $self = shift;
return $self->_do_timeout('read', @_)
}
sub can_write {
@_ == 1 || @_ == 2 || croak(q/Usage: $handle->can_write([timeout])/);
my $self = shift;
return $self->_do_timeout('write', @_)
}
} # HTTP::Micro::Handle
my $prog = <<'EOP';
BEGIN {
if ( defined &IO::Socket::SSL::CAN_IPV6 ) {
*CAN_IPV6 = \*IO::Socket::SSL::CAN_IPV6;
}
else {
constant->import( CAN_IPV6 => '' );
}
my %const = (
NID_CommonName => 13,
GEN_DNS => 2,
GEN_IPADD => 7,
);
while ( my ($name,$value) = each %const ) {
no strict 'refs';
*{$name} = UNIVERSAL::can( 'Net::SSLeay', $name ) || sub { $value };
}
}
{
use Carp qw(croak);
my %dispatcher = (
issuer => sub { Net::SSLeay::X509_NAME_oneline( Net::SSLeay::X509_get_issuer_name( shift )) },
subject => sub { Net::SSLeay::X509_NAME_oneline( Net::SSLeay::X509_get_subject_name( shift )) },
);
if ( $Net::SSLeay::VERSION >= 1.30 ) {
$dispatcher{commonName} = sub {
my $cn = Net::SSLeay::X509_NAME_get_text_by_NID(
Net::SSLeay::X509_get_subject_name( shift ), NID_CommonName);
$cn =~s{\0$}{}; # work around Bug in Net::SSLeay <1.33
$cn;
}
} else {
$dispatcher{commonName} = sub {
croak "you need at least Net::SSLeay version 1.30 for getting commonName"
}
}
if ( $Net::SSLeay::VERSION >= 1.33 ) {
$dispatcher{subjectAltNames} = sub { Net::SSLeay::X509_get_subjectAltNames( shift ) };
} else {
$dispatcher{subjectAltNames} = sub {
return;
};
}
$dispatcher{authority} = $dispatcher{issuer};
$dispatcher{owner} = $dispatcher{subject};
$dispatcher{cn} = $dispatcher{commonName};
sub _peer_certificate {
my ($self, $field) = @_;
my $ssl = $self->_get_ssl_object or return;
my $cert = ${*$self}{_SSL_certificate}
||= Net::SSLeay::get_peer_certificate($ssl)
or return $self->error("Could not retrieve peer certificate");
if ($field) {
my $sub = $dispatcher{$field} or croak
"invalid argument for peer_certificate, valid are: ".join( " ",keys %dispatcher ).
"\nMaybe you need to upgrade your Net::SSLeay";
return $sub->($cert);
} else {
return $cert
}
}
my %scheme = (
ldap => {
wildcards_in_cn => 0,
wildcards_in_alt => 'leftmost',
check_cn => 'always',
},
http => {
wildcards_in_cn => 'anywhere',
wildcards_in_alt => 'anywhere',
check_cn => 'when_only',
},
smtp => {
wildcards_in_cn => 0,
wildcards_in_alt => 0,
check_cn => 'always'
},
none => {}, # do not check
);
$scheme{www} = $scheme{http}; # alias
$scheme{xmpp} = $scheme{http}; # rfc 3920
$scheme{pop3} = $scheme{ldap}; # rfc 2595
$scheme{imap} = $scheme{ldap}; # rfc 2595
$scheme{acap} = $scheme{ldap}; # rfc 2595
$scheme{nntp} = $scheme{ldap}; # rfc 4642
$scheme{ftp} = $scheme{http}; # rfc 4217
sub _verify_hostname_of_cert {
my $identity = shift;
my $cert = shift;
my $scheme = shift || 'none';
if ( ! ref($scheme) ) {
$scheme = $scheme{$scheme} or croak "scheme $scheme not defined";
}
return 1 if ! %$scheme; # 'none'
my $commonName = $dispatcher{cn}->($cert);
my @altNames = $dispatcher{subjectAltNames}->($cert);
if ( my $sub = $scheme->{callback} ) {
return $sub->($identity,$commonName,@altNames);
}
my $ipn;
if ( CAN_IPV6 and $identity =~m{:} ) {
$ipn = IO::Socket::SSL::inet_pton(IO::Socket::SSL::AF_INET6,$identity)
or croak "'$identity' is not IPv6, but neither IPv4 nor hostname";
} elsif ( $identity =~m{^\d+\.\d+\.\d+\.\d+$} ) {
$ipn = IO::Socket::SSL::inet_aton( $identity ) or croak "'$identity' is not IPv4, but neither IPv6 nor hostname";
} else {
if ( $identity =~m{[^a-zA-Z0-9_.\-]} ) {
$identity =~m{\0} and croak("name '$identity' has \\0 byte");
$identity = IO::Socket::SSL::idn_to_ascii($identity) or
croak "Warning: Given name '$identity' could not be converted to IDNA!";
}
}
my $check_name = sub {
my ($name,$identity,$wtyp) = @_;
$wtyp ||= '';
my $pattern;
if ( $wtyp eq 'anywhere' and $name =~m{^([a-zA-Z0-9_\-]*)\*(.+)} ) {
$pattern = qr{^\Q$1\E[a-zA-Z0-9_\-]*\Q$2\E$}i;
} elsif ( $wtyp eq 'leftmost' and $name =~m{^\*(\..+)$} ) {
$pattern = qr{^[a-zA-Z0-9_\-]*\Q$1\E$}i;
} else {
$pattern = qr{^\Q$name\E$}i;
}
return $identity =~ $pattern;
};
my $alt_dnsNames = 0;
while (@altNames) {
my ($type, $name) = splice (@altNames, 0, 2);
if ( $ipn and $type == GEN_IPADD ) {
return 1 if $ipn eq $name;
} elsif ( ! $ipn and $type == GEN_DNS ) {
$name =~s/\s+$//; $name =~s/^\s+//;
$alt_dnsNames++;
$check_name->($name,$identity,$scheme->{wildcards_in_alt})
and return 1;
}
}
if ( ! $ipn and (
$scheme->{check_cn} eq 'always' or
$scheme->{check_cn} eq 'when_only' and !$alt_dnsNames)) {
$check_name->($commonName,$identity,$scheme->{wildcards_in_cn})
and return 1;
}
return 0; # no match
}
}
EOP
eval { require IO::Socket::SSL };
if ( $INC{"IO/Socket/SSL.pm"} ) {
eval $prog;
die $@ if $@;
}
1;
# ###########################################################################
# End HTTP::Micro package
# ###########################################################################
# ###########################################################################
# VersionCheck package
# This package is a copy without comments from the original. The original
# with comments and its test file can be found in the Bazaar repository at,
# lib/VersionCheck.pm
# t/lib/VersionCheck.t
# See https://launchpad.net/percona-toolkit for more information.
# ###########################################################################
{
package VersionCheck;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
use Data::Dumper;
local $Data::Dumper::Indent = 1;
local $Data::Dumper::Sortkeys = 1;
local $Data::Dumper::Quotekeys = 0;
use Digest::MD5 qw(md5_hex);
use Sys::Hostname qw(hostname);
use File::Basename qw();
use File::Spec;
use FindBin qw();
eval {
require Percona::Toolkit;
require HTTP::Micro;
};
{
my $file = 'percona-version-check';
my $home = $ENV{HOME} || $ENV{HOMEPATH} || $ENV{USERPROFILE} || '.';
my @vc_dirs = (
'/etc/percona',
'/etc/percona-toolkit',
'/tmp',
"$home",
);
sub version_check_file {
foreach my $dir ( @vc_dirs ) {
if ( -d $dir && -w $dir ) {
PTDEBUG && _d('Version check file', $file, 'in', $dir);
return $dir . '/' . $file;
}
}
PTDEBUG && _d('Version check file', $file, 'in', $ENV{PWD});
return $file; # in the CWD
}
}
sub version_check_time_limit {
return 60 * 60 * 24; # one day
}
sub version_check {
my (%args) = @_;
my $instances = $args{instances} || [];
my $instances_to_check;
PTDEBUG && _d('FindBin::Bin:', $FindBin::Bin);
if ( !$args{force} ) {
if ( $FindBin::Bin
&& (-d "$FindBin::Bin/../.bzr" ||
-d "$FindBin::Bin/../../.bzr" ||
-d "$FindBin::Bin/../.git" ||
-d "$FindBin::Bin/../../.git"
)
) {
PTDEBUG && _d("$FindBin::Bin/../.bzr disables --version-check");
return;
}
}
eval {
foreach my $instance ( @$instances ) {
my ($name, $id) = get_instance_id($instance);
$instance->{name} = $name;
$instance->{id} = $id;
}
push @$instances, { name => 'system', id => 0 };
$instances_to_check = get_instances_to_check(
instances => $instances,
vc_file => $args{vc_file}, # testing
now => $args{now}, # testing
);
PTDEBUG && _d(scalar @$instances_to_check, 'instances to check');
return unless @$instances_to_check;
my $protocol = 'https';
eval { require IO::Socket::SSL; };
if ( $EVAL_ERROR ) {
PTDEBUG && _d($EVAL_ERROR);
PTDEBUG && _d("SSL not available, won't run version_check");
return;
}
PTDEBUG && _d('Using', $protocol);
my $advice = pingback(
instances => $instances_to_check,
protocol => $protocol,
url => $args{url} # testing
|| $ENV{PERCONA_VERSION_CHECK_URL} # testing
|| "$protocol://v.percona.com",
);
if ( $advice ) {
PTDEBUG && _d('Advice:', Dumper($advice));
if ( scalar @$advice > 1) {
print "\n# " . scalar @$advice . " software updates are "
. "available:\n";
}
else {
print "\n# A software update is available:\n";
}
print join("\n", map { "# * $_" } @$advice), "\n\n";
}
};
if ( $EVAL_ERROR ) {
PTDEBUG && _d('Version check failed:', $EVAL_ERROR);
}
if ( @$instances_to_check ) {
eval {
update_check_times(
instances => $instances_to_check,
vc_file => $args{vc_file}, # testing
now => $args{now}, # testing
);
};
if ( $EVAL_ERROR ) {
PTDEBUG && _d('Error updating version check file:', $EVAL_ERROR);
}
}
if ( $ENV{PTDEBUG_VERSION_CHECK} ) {
warn "Exiting because the PTDEBUG_VERSION_CHECK "
. "environment variable is defined.\n";
exit 255;
}
return;
}
sub get_instances_to_check {
my (%args) = @_;
my $instances = $args{instances};
my $now = $args{now} || int(time);
my $vc_file = $args{vc_file} || version_check_file();
if ( !-f $vc_file ) {
PTDEBUG && _d('Version check file', $vc_file, 'does not exist;',
'version checking all instances');
return $instances;
}
open my $fh, '<', $vc_file or die "Cannot open $vc_file: $OS_ERROR";
chomp(my $file_contents = do { local $/ = undef; <$fh> });
PTDEBUG && _d('Version check file', $vc_file, 'contents:', $file_contents);
close $fh;
my %last_check_time_for = $file_contents =~ /^([^,]+),(.+)$/mg;
my $check_time_limit = version_check_time_limit();
my @instances_to_check;
foreach my $instance ( @$instances ) {
my $last_check_time = $last_check_time_for{ $instance->{id} };
PTDEBUG && _d('Intsance', $instance->{id}, 'last checked',
$last_check_time, 'now', $now, 'diff', $now - ($last_check_time || 0),
'hours until next check',
sprintf '%.2f',
($check_time_limit - ($now - ($last_check_time || 0))) / 3600);
if ( !defined $last_check_time
|| ($now - $last_check_time) >= $check_time_limit ) {
PTDEBUG && _d('Time to check', Dumper($instance));
push @instances_to_check, $instance;
}
}
return \@instances_to_check;
}
sub update_check_times {
my (%args) = @_;
my $instances = $args{instances};
my $now = $args{now} || int(time);
my $vc_file = $args{vc_file} || version_check_file();
PTDEBUG && _d('Updating last check time:', $now);
my %all_instances = map {
$_->{id} => { name => $_->{name}, ts => $now }
} @$instances;
if ( -f $vc_file ) {
open my $fh, '<', $vc_file or die "Cannot read $vc_file: $OS_ERROR";
my $contents = do { local $/ = undef; <$fh> };
close $fh;
foreach my $line ( split("\n", ($contents || '')) ) {
my ($id, $ts) = split(',', $line);
if ( !exists $all_instances{$id} ) {
$all_instances{$id} = { ts => $ts }; # original ts, not updated
}
}
}
open my $fh, '>', $vc_file or die "Cannot write to $vc_file: $OS_ERROR";
foreach my $id ( sort keys %all_instances ) {
PTDEBUG && _d('Updated:', $id, Dumper($all_instances{$id}));
print { $fh } $id . ',' . $all_instances{$id}->{ts} . "\n";
}
close $fh;
return;
}
sub get_instance_id {
my ($instance) = @_;
my $dbh = $instance->{dbh};
my $dsn = $instance->{dsn};
my $sql = q{SELECT CONCAT(@@hostname, @@port)};
PTDEBUG && _d($sql);
my ($name) = eval { $dbh->selectrow_array($sql) };
if ( $EVAL_ERROR ) {
PTDEBUG && _d($EVAL_ERROR);
$sql = q{SELECT @@hostname};
PTDEBUG && _d($sql);
($name) = eval { $dbh->selectrow_array($sql) };
if ( $EVAL_ERROR ) {
PTDEBUG && _d($EVAL_ERROR);
$name = ($dsn->{h} || 'localhost') . ($dsn->{P} || 3306);
}
else {
$sql = q{SHOW VARIABLES LIKE 'port'};
PTDEBUG && _d($sql);
my (undef, $port) = eval { $dbh->selectrow_array($sql) };
PTDEBUG && _d('port:', $port);
$name .= $port || '';
}
}
my $id = md5_hex($name);
PTDEBUG && _d('MySQL instance:', $id, $name, Dumper($dsn));
return $name, $id;
}
sub pingback {
my (%args) = @_;
my @required_args = qw(url instances);
foreach my $arg ( @required_args ) {
die "I need a $arg arugment" unless $args{$arg};
}
my $url = $args{url};
my $instances = $args{instances};
my $ua = $args{ua} || HTTP::Micro->new( timeout => 3 );
my $response = $ua->request('GET', $url);
PTDEBUG && _d('Server response:', Dumper($response));
die "No response from GET $url"
if !$response;
die("GET on $url returned HTTP status $response->{status}; expected 200\n",
($response->{content} || '')) if $response->{status} != 200;
die("GET on $url did not return any programs to check")
if !$response->{content};
my $items = parse_server_response(
response => $response->{content}
);
die "Failed to parse server requested programs: $response->{content}"
if !scalar keys %$items;
my $versions = get_versions(
items => $items,
instances => $instances,
);
die "Failed to get any program versions; should have at least gotten Perl"
if !scalar keys %$versions;
my $client_content = encode_client_response(
items => $items,
versions => $versions,
general_id => md5_hex( hostname() ),
);
my $client_response = {
headers => { "X-Percona-Toolkit-Tool" => File::Basename::basename($0) },
content => $client_content,
};
PTDEBUG && _d('Client response:', Dumper($client_response));
$response = $ua->request('POST', $url, $client_response);
PTDEBUG && _d('Server suggestions:', Dumper($response));
die "No response from POST $url $client_response"
if !$response;
die "POST $url returned HTTP status $response->{status}; expected 200"
if $response->{status} != 200;
return unless $response->{content};
$items = parse_server_response(
response => $response->{content},
split_vars => 0,
);
die "Failed to parse server suggestions: $response->{content}"
if !scalar keys %$items;
my @suggestions = map { $_->{vars} }
sort { $a->{item} cmp $b->{item} }
values %$items;
return \@suggestions;
}
sub encode_client_response {
my (%args) = @_;
my @required_args = qw(items versions general_id);
foreach my $arg ( @required_args ) {
die "I need a $arg arugment" unless $args{$arg};
}
my ($items, $versions, $general_id) = @args{@required_args};
my @lines;
foreach my $item ( sort keys %$items ) {
next unless exists $versions->{$item};
if ( ref($versions->{$item}) eq 'HASH' ) {
my $mysql_versions = $versions->{$item};
for my $id ( sort keys %$mysql_versions ) {
push @lines, join(';', $id, $item, $mysql_versions->{$id});
}
}
else {
push @lines, join(';', $general_id, $item, $versions->{$item});
}
}
my $client_response = join("\n", @lines) . "\n";
return $client_response;
}
sub parse_server_response {
my (%args) = @_;
my @required_args = qw(response);
foreach my $arg ( @required_args ) {
die "I need a $arg arugment" unless $args{$arg};
}
my ($response) = @args{@required_args};
my %items = map {
my ($item, $type, $vars) = split(";", $_);
if ( !defined $args{split_vars} || $args{split_vars} ) {
$vars = [ split(",", ($vars || '')) ];
}
$item => {
item => $item,
type => $type,
vars => $vars,
};
} split("\n", $response);
PTDEBUG && _d('Items:', Dumper(\%items));
return \%items;
}
my %sub_for_type = (
os_version => \&get_os_version,
perl_version => \&get_perl_version,
perl_module_version => \&get_perl_module_version,
mysql_variable => \&get_mysql_variable,
);
sub valid_item {
my ($item) = @_;
return unless $item;
if ( !exists $sub_for_type{ $item->{type} } ) {
PTDEBUG && _d('Invalid type:', $item->{type});
return 0;
}
return 1;
}
sub get_versions {
my (%args) = @_;
my @required_args = qw(items);
foreach my $arg ( @required_args ) {
die "I need a $arg arugment" unless $args{$arg};
}
my ($items) = @args{@required_args};
my %versions;
foreach my $item ( values %$items ) {
next unless valid_item($item);
eval {
my $version = $sub_for_type{ $item->{type} }->(
item => $item,
instances => $args{instances},
);
if ( $version ) {
chomp $version unless ref($version);
$versions{$item->{item}} = $version;
}
};
if ( $EVAL_ERROR ) {
PTDEBUG && _d('Error getting version for', Dumper($item), $EVAL_ERROR);
}
}
return \%versions;
}
sub get_os_version {
if ( $OSNAME eq 'MSWin32' ) {
require Win32;
return Win32::GetOSDisplayName();
}
chomp(my $platform = `uname -s`);
PTDEBUG && _d('platform:', $platform);
return $OSNAME unless $platform;
chomp(my $lsb_release
= `which lsb_release 2>/dev/null | awk '{print \$1}'` || '');
PTDEBUG && _d('lsb_release:', $lsb_release);
my $release = "";
if ( $platform eq 'Linux' ) {
if ( -f "/etc/fedora-release" ) {
$release = `cat /etc/fedora-release`;
}
elsif ( -f "/etc/redhat-release" ) {
$release = `cat /etc/redhat-release`;
}
elsif ( -f "/etc/system-release" ) {
$release = `cat /etc/system-release`;
}
elsif ( $lsb_release ) {
$release = `$lsb_release -ds`;
}
elsif ( -f "/etc/lsb-release" ) {
$release = `grep DISTRIB_DESCRIPTION /etc/lsb-release`;
$release =~ s/^\w+="([^"]+)".+/$1/;
}
elsif ( -f "/etc/debian_version" ) {
chomp(my $rel = `cat /etc/debian_version`);
$release = "Debian $rel";
if ( -f "/etc/apt/sources.list" ) {
chomp(my $code_name = `awk '/^deb/ {print \$3}' /etc/apt/sources.list | awk -F/ '{print \$1}'| awk 'BEGIN {FS="|"} {print \$1}' | sort | uniq -c | sort -rn | head -n1 | awk '{print \$2}'`);
$release .= " ($code_name)" if $code_name;
}
}
elsif ( -f "/etc/os-release" ) { # openSUSE
chomp($release = `grep PRETTY_NAME /etc/os-release`);
$release =~ s/^PRETTY_NAME="(.+)"$/$1/;
}
elsif ( `ls /etc/*release 2>/dev/null` ) {
if ( `grep DISTRIB_DESCRIPTION /etc/*release 2>/dev/null` ) {
$release = `grep DISTRIB_DESCRIPTION /etc/*release | head -n1`;
}
else {
$release = `cat /etc/*release | head -n1`;
}
}
}
elsif ( $platform =~ m/(?:BSD|^Darwin)$/ ) {
my $rel = `uname -r`;
$release = "$platform $rel";
}
elsif ( $platform eq "SunOS" ) {
my $rel = `head -n1 /etc/release` || `uname -r`;
$release = "$platform $rel";
}
if ( !$release ) {
PTDEBUG && _d('Failed to get the release, using platform');
$release = $platform;
}
chomp($release);
$release =~ s/^"|"$//g;
PTDEBUG && _d('OS version =', $release);
return $release;
}
sub get_perl_version {
my (%args) = @_;
my $item = $args{item};
return unless $item;
my $version = sprintf '%vd', $PERL_VERSION;
PTDEBUG && _d('Perl version', $version);
return $version;
}
sub get_perl_module_version {
my (%args) = @_;
my $item = $args{item};
return unless $item;
my $var = '$' . $item->{item} . '::VERSION';
my $version = eval "use $item->{item}; $var;";
PTDEBUG && _d('Perl version for', $var, '=', $version);
return $version;
}
sub get_mysql_variable {
return get_from_mysql(
show => 'VARIABLES',
@_,
);
}
sub get_from_mysql {
my (%args) = @_;
my $show = $args{show};
my $item = $args{item};
my $instances = $args{instances};
return unless $show && $item;
if ( !$instances || !@$instances ) {
PTDEBUG && _d('Cannot check', $item,
'because there are no MySQL instances');
return;
}
if ($item->{item} eq 'MySQL' && $item->{type} eq 'mysql_variable') {
@{$item->{vars}} = grep { $_ eq 'version' || $_ eq 'version_comment' } @{$item->{vars}};
}
my @versions;
my %version_for;
foreach my $instance ( @$instances ) {
next unless $instance->{id}; # special system instance has id=0
my $dbh = $instance->{dbh};
local $dbh->{FetchHashKeyName} = 'NAME_lc';
my $sql = qq/SHOW $show/;
PTDEBUG && _d($sql);
my $rows = $dbh->selectall_hashref($sql, 'variable_name');
my @versions;
foreach my $var ( @{$item->{vars}} ) {
$var = lc($var);
my $version = $rows->{$var}->{value};
PTDEBUG && _d('MySQL version for', $item->{item}, '=', $version,
'on', $instance->{name});
push @versions, $version;
}
$version_for{ $instance->{id} } = join(' ', @versions);
}
return \%version_for;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
1;
}
# ###########################################################################
# End VersionCheck package
# ###########################################################################
# #############################################################################
# This is a combination of modules and programs in one -- a runnable module.
# http://www.perl.com/pub/a/2006/07/13/lightning-articles.html?page=last
# Or, look it up in the Camel book on pages 642 and 643 in the 3rd edition.
#
# Check at the end of this package for the call to main() which actually runs
# the program.
# #############################################################################
package pt_duplicate_key_checker;
use strict;
use warnings FATAL => 'all';
use English qw(-no_match_vars);
use List::Util qw(max);
use Percona::Toolkit;
use constant PTDEBUG => $ENV{PTDEBUG} || 0;
$OUTPUT_AUTOFLUSH = 1;
my $max_width = 74;
my $hdr_width = $max_width - 2; # for '# '
my $hdr_fmt = "# %-${hdr_width}s\n";
sub main {
local @ARGV = @_; # set global ARGV for this package
my %summary = ( 'items' => {'Total Indexes' => 0} );
my %seen_tbl;
my $q = new Quoter();
my $tp = new TableParser(Quoter => $q);
# #######################################################################
# Get configuration information and parse command line options.
# #######################################################################
my $o = new OptionParser();
$o->get_specs();
$o->get_opts();
my $dp = $o->DSNParser();
$dp->prop('set-vars', $o->set_vars());
$o->usage_or_errors();
# ########################################################################
# If --pid, check it first since we'll die if it already exits.
# ########################################################################
my $daemon = Daemon->new(
daemonize => 0,
pid_file => $o->get('pid'),
);
$daemon->run();
# #######################################################################
# Connect to MySQL.
# #######################################################################
if ( $o->got('ask-pass') ) {
$o->set('password', OptionParser::prompt_noecho("Enter password: "));
}
my $dsn_defaults = $dp->parse_options($o);
my $dsn = @ARGV ? $dp->parse(shift @ARGV, $dsn_defaults)
: $dsn_defaults;
my $dbh = $dp->get_dbh($dp->get_cxn_params($dsn),
{ AutoCommit => 1, });
# ########################################################################
# Do the version-check
# ########################################################################
if ( $o->get('version-check') && (!$o->has('quiet') || !$o->get('quiet')) ) {
VersionCheck::version_check(
force => $o->got('version-check'),
instances => [ {dbh => $dbh, dsn => $dsn} ],
);
}
# #######################################################################
# Do the main work.
# #######################################################################
my $ks = $o->get('summary') ? new KeySize(q=>$q) : undef;
my $dk = new DuplicateKeyFinder();
my %tp_opts = (
ignore_type => $o->get('all-structs'),
ignore_order => $o->get('ignore-order'),
clustered => $o->get('clustered'),
);
my $get_keys = $o->get('key-types') =~ m/k/ ? 1 : 0;
my $get_fks = $o->get('key-types') =~ m/f/ ? 1 : 0;
my $schema = new Schema();
my $schema_itr = new SchemaIterator(
dbh => $dbh,
OptionParser => $o,
Quoter => $q,
TableParser => $tp,
Schema => $schema,
);
TABLE:
while ( my $tbl = $schema_itr->next() ) {
eval {
$tbl->{engine} = $tbl->{tbl_struct}->{engine};
my ($keys, $clustered_key, $fks);
if ( $get_keys ) {
($keys, $clustered_key)
= $tp->get_keys($tbl->{ddl}, {});
}
if ( $get_fks ) {
$fks = $tp->get_fks($tbl->{ddl}, {database => $tbl->{db}});
}
if ( ($keys && %$keys) || ($fks && %$fks) ) {
if ( $o->got('verbose') ) {
print_all_keys($keys, $tbl, \%seen_tbl) if $keys;
print_all_keys($fks, $tbl, \%seen_tbl) if $fks;
}
PTDEBUG && _d('Getting duplicate keys on',
$tbl->{db}, $tbl->{tbl});
if ( $keys ) {
$dk->get_duplicate_keys(
$keys,
clustered_key => $clustered_key,
tbl_info => $tbl,
callback => \&print_duplicate_key,
%tp_opts,
# get_duplicate_keys() ignores these args but passes them
# to the callback:
dbh => $dbh,
is_fk => 0,
o => $o,
ks => $ks,
tp => $tp,
q => $q,
seen_tbl => \%seen_tbl,
summary => \%summary,
);
}
if ( $fks ) {
$dk->get_duplicate_fks(
$fks,
tbl_info => $tbl,
callback => \&print_duplicate_key,
%tp_opts,
# get_duplicate_fks() ignores these args but passes them
# to the callback:
dbh => $dbh,
is_fk => 1,
o => $o,
ks => $ks,
tp => $tp,
q => $q,
seen_tbl => \%seen_tbl,
summary => \%summary,
);
}
# Always count Total Keys so print_key_summary won't die
# because %summary is empty.
$summary{items}->{'Total Indexes'} += (scalar keys %$keys)
+ (scalar keys %$fks)
}
};
if ( $EVAL_ERROR ) {
warn "Error checking $tbl->{db}.$tbl->{tbl}: $EVAL_ERROR";
}
} # TABLE
print_key_summary(%summary) if $o->get('summary');
return 0;
}
# ##########################################################################
# Subroutines
# ##########################################################################
sub print_all_keys {
my ( $keys, $tbl_info, $seen_tbl ) = @_;
return unless $keys;
my $db = $tbl_info->{db};
my $tbl = $tbl_info->{tbl};
if ( !$seen_tbl->{"$db$tbl"}++ ) {
printf $hdr_fmt, ('#' x $hdr_width);
printf $hdr_fmt, "$db.$tbl";
printf $hdr_fmt, ('#' x $hdr_width);
}
# Print keys sorted by name (easier to test)
foreach my $key ( sort {$a->{name} cmp $b->{name}} values %$keys ) {
print "\n# $key->{name} ($key->{colnames})";
}
print "\n";
return;
}
sub print_duplicate_key {
my ( $dupe, %args ) = @_;
return unless $dupe;
foreach my $arg ( qw(tbl_info dbh is_fk o ks q tp seen_tbl) ) {
die "I need a $arg argument" unless exists $args{$arg};
}
PTDEBUG && _d('Printing duplicate key', $dupe->{key});
my $db = $args{tbl_info}->{db};
my $tbl = $args{tbl_info}->{tbl};
my $dbh = $args{dbh};
my $o = $args{o};
my $ks = $args{ks};
my $seen_tbl = $args{seen_tbl};
my $q = $args{q};
my $tp = $args{tp};
my $summary = $args{summary};
my $struct = $tp->parse($args{tbl_info}->{ddl});
if ($dupe->{ddl} =~ /FULLTEXT/) {
$summary->{has_fulltext_dupe}++;
}
if ( !$seen_tbl->{"$db$tbl"}++ ) {
printf $hdr_fmt, ('#' x $hdr_width);
printf $hdr_fmt, "$db.$tbl";
printf $hdr_fmt, ('#' x $hdr_width);
print "\n";
}
$dupe->{reason} =~ s/\n/\n# /g;
print "# $dupe->{reason}\n";
print "# Key definitions:\n";
print "# " . ($dupe->{ddl} || '') . "\n";
print "# " . ($dupe->{duplicate_of_ddl} || '') . "\n";
print "# Column types:\n";
my %seen; # print each column only once
foreach my $col ( @{$dupe->{cols}}, @{$dupe->{duplicate_of_cols}} ) {
next if $seen{$col}++;
PTDEBUG && _d('col', $col);
print "#\t" . lc($struct->{defs}->{lc $col}) . "\n";
}
if ( $o->get('sql') ) {
if ( $dupe->{dupe_type} ne 'clustered' ) {
print "# To remove this duplicate "
. ($args{is_fk} ? 'foreign key' : 'index')
. ", execute:\n"
. 'ALTER TABLE ' . $q->quote($db, $tbl)
. ($args{is_fk} ? ' DROP FOREIGN KEY ' : ' DROP INDEX ')
. "`$dupe->{key}`;\n";
}
else {
# Suggest shortening clustered dupes instead of
# removing them (issue 295).
print "# To shorten this duplicate clustered index, execute:\n"
. 'ALTER TABLE '.$q->quote($db, $tbl)." DROP INDEX `$dupe->{key}`, "
. "ADD INDEX `$dupe->{key}` ($dupe->{short_key});\n";
}
}
print "\n";
if ( $o->get('summary') && $summary ) {
$summary->{'items'}->{'Total Duplicate Indexes'} += 1;
my ($size, $chosen_key) = $ks->get_key_size(
name => $dupe->{key},
cols => $dupe->{cols},
tbl_name => $q->quote($db, $tbl),
tbl_struct => $struct,
dbh => $dbh,
);
if ( $args{is_fk} ) {
# Foreign keys have no size because they're just constraints.
print "# MySQL uses the $chosen_key index for this "
. "foreign key constraint\n\n";
}
else {
$size ||= 0;
# Create Size Duplicate Keys summary even if there's no valid keys.
$summary->{'items'}->{'Size Duplicate Indexes'} += $size;
if ( $size ) {
if ( $chosen_key && $chosen_key ne $dupe->{key} ) {
# This shouldn't happen. But in case it does, we should know.
print "# MySQL chose the $chosen_key index despite FORCE INDEX\n\n";
}
}
}
}
return;
}
sub print_key_summary {
my ( %summary ) = @_;
my $items = $summary{items};
printf $hdr_fmt, ('#' x $hdr_width);
printf $hdr_fmt, 'Summary of indexes';
printf $hdr_fmt, ('#' x $hdr_width);
print "\n";
my $max_item = max(map { length($_) } keys %$items);
my $line_fmt = "# %-${max_item}s %-s";
foreach my $item ( sort keys %$items ) {
printf $line_fmt, $item, $items->{$item};
if ( $item eq 'Size Duplicate Indexes' && $summary{has_fulltext_dupe} ) {
print ' (not including FULLTEXT indexes)';
}
print "\n";
}
return;
}
sub _d {
my ($package, undef, $line) = caller 0;
@_ = map { (my $temp = $_) =~ s/\n/\n# /g; $temp; }
map { defined $_ ? $_ : 'undef' }
@_;
print STDERR "# $package:$line $PID ", join(' ', @_), "\n";
}
# ############################################################################
# Run the program.
# ############################################################################
if ( !caller ) { exit main(@ARGV); }
1; # Because this is a module as well as a script.
# ############################################################################
# Documentation
# ############################################################################
=pod
=head1 NAME
pt-duplicate-key-checker - Find duplicate indexes and foreign keys on MySQL tables.
=head1 SYNOPSIS
Usage: pt-duplicate-key-checker [OPTIONS] [DSN]
pt-duplicate-key-checker examines MySQL tables for duplicate or redundant
indexes and foreign keys. Connection options are read from MySQL option files.
pt-duplicate-key-checker --host host1
=head1 RISKS
Percona Toolkit is mature, proven in the real world, and well tested,
but all database tools can pose a risk to the system and the database
server. Before using this tool, please:
=over
=item * Read the tool's documentation
=item * Review the tool's known L<"BUGS">
=item * Test the tool on a non-production server
=item * Backup your production server and verify the backups
=back
=head1 DESCRIPTION
This program examines the output of SHOW CREATE TABLE on MySQL tables, and if
it finds indexes that cover the same columns as another index in the same
order, or cover an exact leftmost prefix of another index, it prints out
the suspicious indexes. By default, indexes must be of the same type, so a
BTREE index is not a duplicate of a FULLTEXT index, even if they have the same
columns. You can override this.
It also looks for duplicate foreign keys. A duplicate foreign key covers the
same columns as another in the same table, and references the same parent
table.
The output ends with a short summary that includes an estimate of the total
size, in bytes, that the duplicate indexes are using. This is calculated by
multiplying the index length by the number of rows in their respective tables.
=head1 OPTIONS
This tool accepts additional command-line arguments. Refer to the
L<"SYNOPSIS"> and usage information for details.
=over
=item --all-structs
Compare indexes with different structs (BTREE, HASH, etc).
By default this is disabled, because a BTREE index that covers the same columns
as a FULLTEXT index is not really a duplicate, for example.
=item --ask-pass
Prompt for a password when connecting to MySQL.
=item --charset
short form: -A; type: string
Default character set. If the value is utf8, sets Perl's binmode on
STDOUT to utf8, passes the mysql_enable_utf8 option to DBD::mysql, and runs SET
NAMES UTF8 after connecting to MySQL. Any other value sets binmode on STDOUT
without the utf8 layer, and runs SET NAMES after connecting to MySQL.
=item --[no]clustered
default: yes
PK columns appended to secondary key is duplicate.
Detects when a suffix of a secondary key is a leftmost prefix of the primary
key, and treats it as a duplicate key. Only detects this condition on storage
engines whose primary keys are clustered (currently InnoDB and solidDB).
Clustered storage engines append the primary key columns to the leaf nodes of
all secondary keys anyway, so you might consider it redundant to have them
appear in the internal nodes as well. Of course, you may also want them in the
internal nodes, because just having them at the leaf nodes won't help for some
queries. It does help for covering index queries, however.
Here's an example of a key that is considered redundant with this option:
PRIMARY KEY (`a`)
KEY `b` (`b`,`a`)
The use of such indexes is rather subtle. For example, suppose you have the
following query:
SELECT ... WHERE b=1 ORDER BY a;
This query will do a filesort if we remove the index on C<b,a>. But if we
shorten the index on C<b,a> to just C<b> and also remove the ORDER BY, the query
should return the same results.
The tool suggests shortening duplicate clustered keys by dropping the key
and re-adding it without the primary key prefix. The shortened clustered
key may still duplicate another key, but the tool cannot currently detect
when this happens without being ran a second time to re-check the newly
shortened clustered keys. Therefore, if you shorten any duplicate clustered
keys, you should run the tool again.
=item --config
type: Array
Read this comma-separated list of config files; if specified, this must be the
first option on the command line.
=item --databases
short form: -d; type: hash
Check only this comma-separated list of databases.
=item --defaults-file
short form: -F; type: string
Only read mysql options from the given file. You must give an absolute pathname.
=item --engines
short form: -e; type: hash
Check only tables whose storage engine is in this comma-separated list.
=item --help
Show help and exit.
=item --host
short form: -h; type: string
Connect to host.
=item --ignore-databases
type: Hash
Ignore this comma-separated list of databases.
=item --ignore-engines
type: Hash
Ignore this comma-separated list of storage engines.
=item --ignore-order
Ignore index order so KEY(a,b) duplicates KEY(b,a).
=item --ignore-tables
type: Hash
Ignore this comma-separated list of tables. Table names may be qualified with
the database name.
=item --key-types
type: string; default: fk
Check for duplicate f=foreign keys, k=keys or fk=both.
=item --password
short form: -p; type: string
Password to use when connecting.
If password contains commas they must be escaped with a backslash: "exam\,ple"
=item --pid
type: string
Create the given PID file. The tool won't start if the PID file already
exists and the PID it contains is different than the current PID. However,
if the PID file exists and the PID it contains is no longer running, the
tool will overwrite the PID file with the current PID. The PID file is
removed automatically when the tool exits.
=item --port
short form: -P; type: int
Port number to use for connection.
=item --set-vars
type: Array
Set the MySQL variables in this comma-separated list of C<variable=value> pairs.
By default, the tool sets:
=for comment ignore-pt-internal-value
MAGIC_set_vars
wait_timeout=10000
Variables specified on the command line override these defaults. For
example, specifying C<--set-vars wait_timeout=500> overrides the defaultvalue of C<10000>.
The tool prints a warning and continues if a variable cannot be set.
=item --socket
short form: -S; type: string
Socket file to use for connection.
=item --[no]sql
default: yes
Print DROP KEY statement for each duplicate key. By default an ALTER TABLE
DROP KEY statement is printed below each duplicate key so that, if you want to
remove the duplicate key, you can copy-paste the statement into MySQL.
To disable printing these statements, specify C<--no-sql>.
=item --[no]summary
default: yes
Print summary of indexes at end of output.
=item --tables
short form: -t; type: hash
Check only this comma-separated list of tables.
Table names may be qualified with the database name.
=item --user
short form: -u; type: string
User for login if not current user.
=item --verbose
short form: -v
Output all keys and/or foreign keys found, not just redundant ones.
=item --version
Show version and exit.
=item --[no]version-check
default: yes
Check for the latest version of Percona Toolkit, MySQL, and other programs.
This is a standard "check for updates automatically" feature, with two
additional features. First, the tool checks the version of other programs
on the local system in addition to its own version. For example, it checks
the version of every MySQL server it connects to, Perl, and the Perl module
DBD::mysql. Second, it checks for and warns about versions with known
problems. For example, MySQL 5.5.25 had a critical bug and was re-released
as 5.5.25a.
Any updates or known problems are printed to STDOUT before the tool's normal
output. This feature should never interfere with the normal operation of the
tool.
For more information, visit L<https://www.percona.com/version-check>.
=back
=head1 DSN OPTIONS
These DSN options are used to create a DSN. Each option is given like
C<option=value>. The options are case-sensitive, so P and p are not the
same option. There cannot be whitespace before or after the C<=> and
if the value contains whitespace it must be quoted. DSN options are
comma-separated. See the L<percona-toolkit> manpage for full details.
=over
=item * A
dsn: charset; copy: yes
Default character set.
=item * D
dsn: database; copy: yes
Default database.
=item * F
dsn: mysql_read_default_file; copy: yes
Only read default options from the given file
=item * h
dsn: host; copy: yes
Connect to host.
=item * p
dsn: password; copy: yes
Password to use when connecting.
If password contains commas they must be escaped with a backslash: "exam\,ple"
=item * P
dsn: port; copy: yes
Port number to use for connection.
=item * S
dsn: mysql_socket; copy: yes
Socket file to use for connection.
=item * u
dsn: user; copy: yes
User for login if not current user.
=back
=head1 ENVIRONMENT
The environment variable C<PTDEBUG> enables verbose debugging output to STDERR.
To enable debugging and capture all output to a file, run the tool like:
PTDEBUG=1 pt-duplicate-key-checker ... > FILE 2>&1
Be careful: debugging output is voluminous and can generate several megabytes
of output.
=head1 SYSTEM REQUIREMENTS
You need Perl, DBI, DBD::mysql, and some core packages that ought to be
installed in any reasonably new version of Perl.
=head1 BUGS
For a list of known bugs, see L<http://www.percona.com/bugs/pt-duplicate-key-checker>.
Please report bugs at L<https://bugs.launchpad.net/percona-toolkit>.
Include the following information in your bug report:
=over
=item * Complete command-line used to run the tool
=item * Tool L<"--version">
=item * MySQL version of all servers involved
=item * Output from the tool including STDERR
=item * Input files (log/dump/config files, etc.)
=back
If possible, include debugging output by running the tool with C<PTDEBUG>;
see L<"ENVIRONMENT">.
=head1 DOWNLOADING
Visit L<http://www.percona.com/software/percona-toolkit/> to download the
latest release of Percona Toolkit. Or, get the latest release from the
command line:
wget percona.com/get/percona-toolkit.tar.gz
wget percona.com/get/percona-toolkit.rpm
wget percona.com/get/percona-toolkit.deb
You can also get individual tools from the latest release:
wget percona.com/get/TOOL
Replace C<TOOL> with the name of any tool.
=head1 AUTHORS
Baron Schwartz and Daniel Nichter
=head1 ABOUT PERCONA TOOLKIT
This tool is part of Percona Toolkit, a collection of advanced command-line
tools for MySQL developed by Percona. Percona Toolkit was forked from two
projects in June, 2011: Maatkit and Aspersa. Those projects were created by
Baron Schwartz and primarily developed by him and Daniel Nichter. Visit
L<http://www.percona.com/software/> to learn about other free, open-source
software from Percona.
=head1 COPYRIGHT, LICENSE, AND WARRANTY
This program is copyright 2011-2015 Percona LLC and/or its affiliates,
2007-2011 Baron Schwartz.
THIS PROGRAM IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED
WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation, version 2; OR the Perl Artistic License. On UNIX and similar
systems, you can issue `man perlgpl' or `man perlartistic' to read these
licenses.
You should have received a copy of the GNU General Public License along with
this program; if not, write to the Free Software Foundation, Inc., 59 Temple
Place, Suite 330, Boston, MA 02111-1307 USA.
=head1 VERSION
pt-duplicate-key-checker 2.2.16
=cut