1
by brian
clean slate |
1 |
#!/usr/bin/perl
|
2 |
# mysqldumpslow - parse and summarize the MySQL slow query log
|
|
3 |
||
4 |
# Original version by Tim Bunce, sometime in 2000.
|
|
5 |
# Further changes by Tim Bunce, 8th March 2001.
|
|
6 |
# Handling of strings with \ and double '' by Monty 11 Aug 2001.
|
|
7 |
||
8 |
use strict;
|
|
9 |
use Getopt::Long;
|
|
10 |
||
11 |
# t=time, l=lock time, r=rows
|
|
12 |
# at, al, and ar are the corresponding averages
|
|
13 |
||
14 |
my %opt = ( |
|
15 |
s => 'at', |
|
16 |
h => '*', |
|
17 |
); |
|
18 |
||
19 |
GetOptions(\%opt, |
|
20 |
'verbose|v+',# verbose
|
|
21 |
'help+', # write usage info |
|
22 |
'debug|d+', # debug |
|
23 |
's=s', # what to sort by (t, at, l, al, r, ar etc) |
|
24 |
'r!', # reverse the sort order (largest last instead of first) |
|
25 |
't=i', # just show the top n queries |
|
26 |
'a!', # don't abstract all numbers to N and strings to 'S' |
|
27 |
'n=i', # abstract numbers with at least n digits within names |
|
28 |
'g=s', # grep: only consider stmts that include this string |
|
29 |
'h=s', # hostname of db server for *-slow.log filename (can be wildcard) |
|
30 |
'i=s', # name of server instance (if using mysql.server startup script) |
|
31 |
'l!', # don't subtract lock time from total time |
|
32 |
) or usage("bad option"); |
|
33 |
||
34 |
$opt{'help'} and usage(); |
|
35 |
||
36 |
unless (@ARGV) { |
|
37 |
my $defaults = `my_print_defaults mysqld`; |
|
38 |
my $basedir = ($defaults =~ m/--basedir=(.*)/)[0] |
|
39 |
or die "Can't determine basedir from 'my_print_defaults mysqld' output: $defaults"; |
|
40 |
warn "basedir=$basedir\n" if $opt{v}; |
|
41 |
||
42 |
my $datadir = ($defaults =~ m/--datadir=(.*)/)[0]; |
|
43 |
my $slowlog = ($defaults =~ m/--log-slow-queries=(.*)/)[0]; |
|
44 |
if (!$datadir or $opt{i}) { |
|
45 |
# determine the datadir from the instances section of /etc/my.cnf, if any
|
|
46 |
my $instances = `my_print_defaults instances`; |
|
47 |
die "Can't determine datadir from 'my_print_defaults mysqld' output: $defaults" |
|
48 |
unless $instances; |
|
49 |
my @instances = ($instances =~ m/^--(\w+)-/mg); |
|
50 |
die "No -i 'instance_name' specified to select among known instances: @instances.\n"
|
|
51 |
unless $opt{i}; |
|
52 |
die "Instance '$opt{i}' is unknown (known instances: @instances)\n" |
|
53 |
unless grep { $_ eq $opt{i} } @instances; |
|
54 |
$datadir = ($instances =~ m/--$opt{i}-datadir=(.*)/)[0] |
|
55 |
or die "Can't determine --$opt{i}-datadir from 'my_print_defaults instances' output: $instances"; |
|
56 |
warn "datadir=$datadir\n" if $opt{v}; |
|
57 |
}
|
|
58 |
||
59 |
if ( -f $slowlog ) { |
|
60 |
@ARGV = ($slowlog); |
|
61 |
die "Can't find '$slowlog'\n" unless @ARGV; |
|
62 |
} else { |
|
63 |
@ARGV = <$datadir/$opt{h}-slow.log>; |
|
64 |
die "Can't find '$datadir/$opt{h}-slow.log'\n" unless @ARGV; |
|
65 |
}
|
|
66 |
}
|
|
67 |
||
68 |
warn "\nReading mysql slow query log from @ARGV\n"; |
|
69 |
||
70 |
my @pending;
|
|
71 |
my %stmt;
|
|
72 |
$/ = ";\n#"; # read entire statements using paragraph mode |
|
73 |
while ( defined($_ = shift @pending) or defined($_ = <>) ) { |
|
74 |
warn "[[$_]]\n" if $opt{d}; # show raw paragraph being read |
|
75 |
||
76 |
my @chunks = split /^\/.*Version.*started with[\000-\377]*?Time.*Id.*Command.*Argument.*\n/m; |
|
77 |
if (@chunks > 1) { |
|
78 |
unshift @pending, map { length($_) ? $_ : () } @chunks; |
|
79 |
warn "<<".join(">>\n<<",@chunks).">>" if $opt{d}; |
|
80 |
next;
|
|
81 |
}
|
|
82 |
||
83 |
s/^#? Time: \d{6}\s+\d+:\d+:\d+.*\n//; |
|
84 |
my ($user,$host) = s/^#? User\@Host:\s+(\S+)\s+\@\s+(\S+).*\n// ? ($1,$2) : ('',''); |
|
85 |
||
86 |
s/^# Query_time: (\d+) Lock_time: (\d+) Rows_sent: (\d+).*\n//; |
|
87 |
my ($t, $l, $r) = ($1, $2, $3); |
|
88 |
$t -= $l unless $opt{l}; |
|
89 |
||
90 |
# remove fluff that mysqld writes to log when it (re)starts:
|
|
91 |
s!^/.*Version.*started with:.*\n!!mg; |
|
92 |
s!^Tcp port: \d+ Unix socket: \S+\n!!mg; |
|
93 |
s!^Time.*Id.*Command.*Argument.*\n!!mg; |
|
94 |
||
95 |
s/^use \w+;\n//; # not consistently added |
|
96 |
s/^SET timestamp=\d+;\n//; |
|
97 |
||
98 |
s/^[ ]*\n//mg; # delete blank lines |
|
99 |
s/^[ ]*/ /mg; # normalize leading whitespace |
|
100 |
s/\s*;\s*(#\s*)?$//; # remove trailing semicolon(+newline-hash) |
|
101 |
||
102 |
next if $opt{g} and !m/$opt{g}/io; |
|
103 |
||
104 |
unless ($opt{a}) { |
|
105 |
s/\b\d+\b/N/g; |
|
106 |
s/\b0x[0-9A-Fa-f]+\b/N/g; |
|
107 |
s/''/'S'/g; |
|
108 |
s/""/"S"/g; |
|
109 |
s/(\\')//g; |
|
110 |
s/(\\")//g;
|
|
111 |
s/'[^']+'/'S'/g; |
|
112 |
s/"[^"]+"/"S"/g; |
|
113 |
# -n=8: turn log_20001231 into log_NNNNNNNN
|
|
114 |
s/([a-z_]+)(\d{$opt{n},})/$1.('N' x length($2))/ieg if $opt{n}; |
|
115 |
# abbreviate massive "in (...)" statements and similar |
|
116 |
s!(([NS],){100,})!sprintf("$2,{repeated %d times}",length($1)/2)!eg; |
|
117 |
}
|
|
118 |
||
119 |
my $s = $stmt{$_} ||= { users=>{}, hosts=>{} }; |
|
120 |
$s->{c} += 1; |
|
121 |
$s->{t} += $t; |
|
122 |
$s->{l} += $l; |
|
123 |
$s->{r} += $r; |
|
124 |
$s->{users}->{$user}++ if $user; |
|
125 |
$s->{hosts}->{$host}++ if $host; |
|
126 |
||
127 |
warn "{{$_}}\n\n" if $opt{d}; # show processed statement string |
|
128 |
}
|
|
129 |
||
130 |
foreach (keys %stmt) {
|
|
131 |
my $v = $stmt{$_} || die; |
|
132 |
my ($c, $t, $l, $r) = @{ $v }{qw(c t l r)}; |
|
133 |
$v->{at} = $t / $c; |
|
134 |
$v->{al} = $l / $c; |
|
135 |
$v->{ar} = $r / $c; |
|
136 |
}
|
|
137 |
||
138 |
my @sorted = sort { $stmt{$b}->{$opt{s}} <=> $stmt{$a}->{$opt{s}} } keys %stmt; |
|
139 |
@sorted = @sorted[0 .. $opt{t}-1] if $opt{t}; |
|
140 |
@sorted = reverse @sorted if $opt{r}; |
|
141 |
||
142 |
foreach (@sorted) {
|
|
143 |
my $v = $stmt{$_} || die; |
|
144 |
my ($c, $t,$at, $l,$al, $r,$ar) = @{ $v }{qw(c t at l al r ar)}; |
|
145 |
my @users = keys %{$v->{users}}; |
|
146 |
my $user = (@users==1) ? $users[0] : sprintf "%dusers",scalar @users; |
|
147 |
my @hosts = keys %{$v->{hosts}}; |
|
148 |
my $host = (@hosts==1) ? $hosts[0] : sprintf "%dhosts",scalar @hosts; |
|
149 |
printf "Count: %d Time=%.2fs (%ds) Lock=%.2fs (%ds) Rows=%.1f (%d), $user\@$host\n%s\n\n", |
|
150 |
$c, $at,$t, $al,$l, $ar,$r, $_; |
|
151 |
}
|
|
152 |
||
153 |
sub usage {
|
|
154 |
my $str= shift; |
|
155 |
my $text= <<HERE; |
|
156 |
Usage: mysqldumpslow [ OPTS... ] [ LOGS... ]
|
|
157 |
||
158 |
Parse and summarize the MySQL slow query log. Options are
|
|
159 |
||
160 |
--verbose verbose
|
|
161 |
--debug debug
|
|
162 |
--help write this text to standard output
|
|
163 |
||
164 |
-v verbose
|
|
165 |
-d debug
|
|
166 |
-s ORDER what to sort by (t, at, l, al, r, ar etc), 'at' is default
|
|
167 |
-r reverse the sort order (largest last instead of first)
|
|
168 |
-t NUM just show the top n queries
|
|
169 |
-a don't abstract all numbers to N and strings to 'S'
|
|
170 |
-n NUM abstract numbers with at least n digits within names
|
|
171 |
-g PATTERN grep: only consider stmts that include this string
|
|
172 |
-h HOSTNAME hostname of db server for *-slow.log filename (can be wildcard),
|
|
173 |
default is '*', i.e. match all
|
|
174 |
-i NAME name of server instance (if using mysql.server startup script)
|
|
175 |
-l don't subtract lock time from total time
|
|
176 |
||
177 |
HERE
|
|
178 |
if ($str) { |
|
179 |
print STDERR "ERROR: $str\n\n"; |
|
180 |
print STDERR $text; |
|
181 |
exit 1;
|
|
182 |
} else {
|
|
183 |
print $text; |
|
184 |
exit 0;
|
|
185 |
}
|
|
186 |
}
|