=================================================================== RCS file: /cvs/nagios/check_hw_sensors/check_hw_sensors,v retrieving revision 1.14 retrieving revision 1.25 diff -u -r1.14 -r1.25 --- nagios/check_hw_sensors/check_hw_sensors 2006/05/04 02:30:29 1.14 +++ nagios/check_hw_sensors/check_hw_sensors 2008/03/10 17:21:53 1.25 @@ -1,5 +1,5 @@ #!/usr/bin/perl -T -# $RedRiver: check_hw_sensors,v 1.13 2006/05/03 22:16:42 andrew Exp $ +# $RedRiver: check_hw_sensors,v 1.24 2007/02/14 21:59:10 andrew Exp $ ######################################################################## # check_hw_sensors *** A nagios check for OpenBSD hw.sensors # @@ -7,21 +7,17 @@ ######################################################################## # TODO: # Really need real documentation. -# -# I want the ability to just check the "status" entry that is in -# some output. For example the OK here: -# hw.sensors.1=esm0, CPU 1, OK, temp, 31.00 degC / 87.80 degF +# allow checking of hw.sensors on a remote host with ssh somehow ######################################################################## use strict; use warnings; -#use Data::Dumper; - %ENV = (); use constant NAGIOS_OUTPUT => 1; use POSIX; +use Config; use lib "/usr/local/libexec/nagios"; use utils qw($TIMEOUT %ERRORS &print_revision &support); @@ -34,10 +30,12 @@ my $GETCAP = '/usr/bin/getcap'; my $BASE = 'hw.sensors'; my $DEFAULT_CONFIG = '/etc/sensorsd.conf'; +my $OSVer = $Config{'osvers'} || 0; my $state = 'UNKNOWN'; # tells whether the it is warning, critical, or OK my %states; # This stores the count of states; my $filename; +my $ignore_status; my $sensor; my $warning; my $critical; @@ -50,34 +48,84 @@ #Option checking my $status = GetOptions( - "version|V" => \$opt_V, - "help|h" => \$opt_h, - "filename|f:s" => \$filename, - "sensor|s=s" => \$sensor, - "warning|w=s" => \$warning, - "critical|c=s" => \$critical, + "version|V" => \$opt_V, + "filename|f:s" => \$filename, + "ignore-status|i" => \$ignore_status, + "sensor|s=s" => \$sensor, + "warning|w=s" => \$warning, + "critical|c=s" => \$critical, ); - + # set the default this way so it only happens if someone typed -f or --filename $filename = $DEFAULT_CONFIG if (defined $filename && $filename eq ''); +# Stuff is output in this file by print_sensor() +# http://www.openbsd.org/cgi-bin/cvsweb/src/sbin/sysctl/sysctl.c +my @Type_Map = ( + { + type => 'temp', + regex => qr/\sdegC$/, + }, + { + type => 'fanrpm', + regex => qr/\sRPM$/, + }, + { + type => 'volts_dc', + regex => qr/\sV\sDC$/, + }, + { + type => 'amps', + regex => qr/\sA$/, + }, + { + type => 'watthour', + regex => qr/\sWh$/, + }, + { + type => 'amphour', + regex => qr/\sAh$/, + }, + { + type => 'indicator', + regex => qr/^(On|Off)$/, + }, + { + type => 'integer', + regex => qr/\sraw$/, + }, + { + type => 'percent', + regex => qr/\s\%$/, + }, + { + type => 'lux', + regex => qr/\slx$/, + }, + { + type => 'drive', + regex => qr/^drive\s/, + }, + { + type => 'timedelta', + regex => qr/\ssecs$/, + }, +); + if ($status == 0) { print_help() ; exit $ERRORS{'OK'}; } if ($opt_V) { - print_revision($PROGNAME,'$Revision: 1.14 $ '); + print_revision($PROGNAME,'$Revision: 1.25 $ '); exit $ERRORS{'OK'}; } -if ($opt_h) { - print_help(); - exit $ERRORS{'OK'}; -} - -unless ( ( +unless ( + ( defined $filename || + (not defined $ignore_status) || (defined $sensor && ($warning || $critical)) ) && ( (!defined $filename) || (!defined $sensor) ) @@ -93,10 +141,9 @@ } $CHECK_SENSOR = $sensor; - $CHECKS{$sensor} = { - 'warn' => $warning, - 'crit' => $critical, - }; + $CHECKS{$sensor}{'warn'} = $warning if defined $warning; + $CHECKS{$sensor}{'crit'} = $critical if defined $critical; + } elsif (defined $filename) { %CHECKS = parse_file($filename); } @@ -107,14 +154,38 @@ #while (<>) { chomp; my ($id, $output) = split /=/; + my @s = split /\./, $id; my @o = split /,\s*/, $output; - my $source = $o[0]; - my $descr = $o[1]; - my $status = $o[2] if @o == 5; - my $type = $o[-2]; - my $data = $o[-1]; + my ($type, $source, $descr, $data, $status); + $source = $o[0]; + $descr = $o[1]; + + if ($OSVer >= 4.1) { + $data = $o[0]; + if ($data =~ s/\s+\((.*)\).*$//) { + $descr = $1; + } + $status = $o[1]; + ($source, $type) = $id =~ /([^\.]+)\.([^\.]+?)\d+$/; + } elsif ($OSVer >= 4.0) { + $data = $o[2]; + $status = $o[3]; + foreach my $t (@Type_Map) { + if ($data =~ /$t->{'regex'}/) { + $type = $t->{'type'}; + last; + } + } + } else { + $data = $o[-1]; + $status = $o[2] if @o == 5; + $type = $o[-2]; + } + + $type ||= 'unknown'; + $SENSORS{$id} = { id => $id, output => $output, @@ -136,18 +207,21 @@ $_a <=> $_b; } -foreach my $check (sort as_if_numeric keys %CHECKS) { - if (exists $SENSORS{$check}) { - my $r = check_sensor($CHECKS{$check}, $SENSORS{$check}); - push @{ $states{ $r } }, - $check . '=' . $SENSORS{$check}{'output'}; +foreach my $s (sort as_if_numeric keys %SENSORS) { + my ($r, $data); + if (exists $CHECKS{$s}) { + $r = check_sensor($SENSORS{$s}, $CHECKS{$s}); + $data = $s . '=' . $SENSORS{$s}{'output'}; + } elsif (not $ignore_status) { + $r = check_sensor($SENSORS{$s}); + $data = $s . '=' . $SENSORS{$s}{'output'}; } else { - push @{ $states{'UNKNOWN'} }, $check . '=No sensor with this id'; + # ignore this sensor } + next unless defined $r; + push @{ $states{ $r } }, $data; } -#print Dumper \%states; - $state = 'OK'; my $have_results = 0; foreach my $error (sort { $ERRORS{$a} <=> $ERRORS{$b} } keys %ERRORS) { @@ -162,7 +236,7 @@ print "$error (" . scalar(@{ $states{ $error } }) . ")"; unless ($error eq 'OK') { print '
'; - print map { " - $_
" } @{ $states{ $error } }; + print map {" - $_
"} @{ $states{ $error } }; } } else { print "$error (" . scalar(@{ $states{ $error } }) . "):\n"; @@ -181,15 +255,23 @@ sub parse_file { my $filename = shift; my %contents; + + die "file '$filename' does not exist." unless -e $filename; open my $fh, '-|', $GETCAP, '-a', '-f', $filename or die "Couldn't open FILE '$GETCAP -a -f $filename': $!"; - while (<$fh>) { + LINE: while (<$fh>) { chomp; my ($key, @c) = split /\:/; foreach (@c) { my ($k, $v) = split /\=/; - $contents{$key}{$k} = $v; + if (lc($k) eq 'ignore') { + $contents{$key}{'IGNORE'} = 1; + } elsif (lc($k) eq 'status') { + $contents{$key}{'STATUS'} = 1; + } else { + $contents{$key}{$k} = $v; + } } } close $fh; @@ -201,6 +283,10 @@ my $type = shift; my $check = shift; + return undef unless $check; + return undef if $check->{'STATUS'}; + return 'IGNORE' if $check->{'IGNORE'}; + foreach my $code ('crit', 'warn') { if (defined $check->{$code} && $check->{$code} =~ /:/) { if (my ($low, $high) = split /:/, $check->{$code}) { @@ -224,8 +310,8 @@ } sub check_sensor { - my $check = shift; my $sensor = shift; + my $check = shift; my $result = 'UNKNOWN'; my %errors = ( 'warn' => 'WARNING', @@ -233,17 +319,29 @@ ); return $result unless ref $sensor eq 'HASH'; + $check = parse_check($sensor->{'type'}, $check) if $check; - $check = parse_check($sensor->{'type'}, $check); - #print Dumper $check, $sensor; + # It looks like doing this should be safe, from + # src/sbin/sysctl/sysctl.c + return $sensor->{'status'} unless $check; - $result = 'OK'; + return undef if $check eq 'IGNORE'; + $result = 'OK'; foreach my $code ('warn', 'crit') { if ( - $sensor->{'type'} eq 'volts_dc' || + $sensor->{'type'} eq 'fan' || $sensor->{'type'} eq 'fanrpm' || - $sensor->{'type'} eq 'raw' + $sensor->{'type'} eq 'volt' || + $sensor->{'type'} eq 'volts_dc' || + $sensor->{'type'} eq 'amps' || + $sensor->{'type'} eq 'watthour' || + $sensor->{'type'} eq 'amphour' || + $sensor->{'type'} eq 'integer' || + $sensor->{'type'} eq 'raw' || + $sensor->{'type'} eq 'percent' || + $sensor->{'type'} eq 'lux' || + $sensor->{'type'} eq 'timedelta' ) { my $data = $sensor->{'data'}; $data =~ s/[^\d\.]//g; @@ -291,7 +389,7 @@ next; } - if ($_ eq $data) { + if ($c eq $data) { $matched = 1; last; } @@ -302,6 +400,7 @@ } elsif ($sensor->{'type'} eq 'temp') { my ($degC, $degF) = split /\//, $sensor->{'data'}; $degC =~ s/[^\d\.]//g; + $degF ||= $degC * 9 / 5 + 32; $degF =~ s/[^\d\.]//g; if ( defined $check->{$code . ".low"} || @@ -369,7 +468,7 @@ next; } - if ($_ eq $data) { + if ($c eq $data) { $matched = 1; last; } @@ -381,6 +480,7 @@ $sensor->{'type'} eq 'drive' || $sensor->{'type'} eq 'indicator' ) { + $sensor->{'data'} =~ s/^drive\s+//; if (@{ $check->{$code} }) { my $matched = 0; foreach (@{ $check->{$code} }) { @@ -393,6 +493,11 @@ } } else { + print STDERR 'Unknown Sensor Type: ', + $sensor->{'id'}, + '=', + $sensor->{'type'}, + "\n"; $result = 'UNKNOWN'; } @@ -404,9 +509,11 @@ sub print_help { print <]|(-s -w limit -c limit)) + $PROGNAME [-i] (-f []|(-s [-w limit] [-c limit])) Usage: + -i, --ignore-status + Don't check the status of sensors that report it. -f, --filename=FILE FILE to load checks from (defaults to /etc/sensorsd.conf) -s, --sensor=ID @@ -416,30 +523,30 @@ -c, --critical=RANGE or single ENTRY Exit with CRITICAL status if outside of RANGE or if != ENTRY - -h (--help) usage help - FILE is in the same format as sensorsd.conf(5) plus some additional entries. These additional entries in the file are ignored by sensorsd(8). $PROGNAME understands the following entries: - low, high, crit, warn, crit.low, crit.high, warn.low, warn.high + low, high, crit, warn, crit.low, crit.high, warn.low, warn.high, + ignore, status An ENTRY depends on the type. The descriptions in sensorsd.conf(5) can be used when appropriate, or you can use the following: - volts_dc, fanrpm or raw - Anything that includes digits. - Both the value of the check and the value of the sensor - response that are not either a digit or period are stripped - and then the two resultant values are compared. + fanrpm, volts_dc, amps, watthour, amphour, integer (raw), percent, + lux or timedelta - Anything that includes digits. Both the value of + the check and the value of the sensor response that are not either a + digit or period are stripped and then the two resultant values are + compared. temp - Can be as above, but if the entry has an F in it, it compares farenheit, otherwise it uses celcius. indicator or drive - does a case sensitive match of each entry in the comma separated list and if it does not match - any of the entries, it matches the status. + any of the entries, it sets the status. The entries 'crit' or 'warn' (or the -c or -w on the command line) may be a RANGE or a comma separated list of acceptable values. @@ -451,8 +558,15 @@ It can also be low: or :high with the other side left blank to only make the single check.. +An entry marked "ignore" will cause that sensor to be skipped. +Generally used with state checking of all sensors to ignore sensors you +don't care about or that report incorrectly. + +If you are using --ignore-status, you can still check the status of +individual sensors with a status entry. + EOL - - print_revision($PROGNAME, '$Revision: 1.14 $'); + + print_revision($PROGNAME, '$Revision: 1.25 $'); }