--- /dev/null
+--- mon-0.99.2.orig/doc/mon.8
++++ mon-0.99.2/doc/mon.8
+@@ -75,7 +75,7 @@
+ Read configuration from
+ .IR file .
+ This defaults to
+-IR /etc/mon/mon.cf " if the " /etc/mon
++.IR /etc/mon/mon.cf " if the " /etc/mon
+ directory exists, otherwise to
+ .IR /etc/mon.cf .
+ .TP
+@@ -241,7 +241,7 @@
+ monitor fping.monitor -t 4000 -r 2
+ \fR
+ .br
+-then the exectuable "fping.monitor" will be executed with these
++then the executable "fping.monitor" will be executed with these
+ parameters:
+ .br
+ \fC
+@@ -382,6 +382,7 @@
+ parameter, or in the
+ .I /usr/local/lib/mon/alert.d
+ and
++.I /usr/lib/mon/alert.d
+ directories if not specified. They are invoked with the following command-line
+ parameters:
+
+@@ -551,8 +552,11 @@
+ .B \-s
+ command-line parameter.
+
+-Multiple alert paths may be specified by separating them with
+-a colon. All paths must be absolute.
++Multiple monitor paths may be specified by separating them with
++a colon. Non-absolute paths are taken to be relative to the
++base directory
++.RI ( /usr/lib/mon
++by default).
+
+ When the configuration file is read, all monitors referenced from the
+ configuration will be looked up in each of these paths, and the
+@@ -909,6 +913,13 @@
+ .IR timeval .
+
+ .TP
++.BI failure_interval " timeval"
++Adjusts the polling interval to
++.I timeval
++when the service check is failing. Resets the interval
++to the original when the service succeeds.
++
++.TP
+ .BI traptimeout " timeval"
+ This keyword takes the same time specification argument as
+ .BI interval ,
+@@ -1206,13 +1217,6 @@
+ automatically to alert programs.
+
+ .TP
+-.BI failure_interval " timeval"
+-Adjusts the polling interval to
+-.I timeval
+-when the service check is failing. Resets the interval
+-to the original when the service succeeds.
+-
+-.TP
+ .BI upalert " alert [arg...]"
+ An
+ .B upalert
+--- mon-0.99.2.orig/mon
++++ mon-0.99.2/mon
+@@ -1369,6 +1369,21 @@
+ $pref->{"no_comp_alerts"} = 1;
+ next;
+ }
++
++ #
++ # for backwards-compatibility with <= 0.38.21
++ #
++ elsif ($var eq "comp_alerts")
++ {
++ $pref->{"no_comp_alerts"} = 0;
++ next;
++ }
++
++ else
++ {
++ close (CFG);
++ return "cf error: unknown syntax in period section [$l], line $line_num";
++ }
+ }
+
+ #
+@@ -1547,10 +1562,13 @@
+ # reap children that may be running
+ #
+ foreach my $pid (keys %runningpid) {
+- my ($group, $service) = split (/\//, $runningpid{$pid});
+ kill 15, $pid;
+- waitpid ($pid, 0);
+- syslog ('info', "reset killed child $pid, exit status $?");
++ # XXX don't wait forever
++ if (waitpid ($pid, 0) == $pid) {
++ syslog ('info', "reset killed child $pid, exit status $?");
++ } else {
++ syslog ('notice', "reset couldn't reap child $pid ($!)");
++ }
+ remove_proc ($pid);
+ }
+
+@@ -1717,7 +1735,7 @@
+ socket (TRAPSERVER, PF_INET, SOCK_DGRAM, $udpproto) ||
+ die_die ("err", "could not create UDP socket: $!");
+ bind (TRAPSERVER, sockaddr_in ($CF{"TRAPPORT"}, $bindaddr)) ||
+- die_die ("err", "could not bind UDP server port: $!");
++ die_die ("err", "could not bind UDP trap port $CF{'TRAPPORT'}: $!");
+ configure_filehandle (*TRAPSERVER) ||
+ die_die ("err", "could not configure UDP trap port: $!");
+
+@@ -1729,7 +1747,7 @@
+ socket (SNMPSERVER, PF_INET, SOCK_DGRAM, $udpproto) ||
+ die_die ("err", "could not create UDP socket: $!");
+ bind (SNMPSERVER, sockaddr_in ($CF{"SNMPPORT"}, INADDR_ANY)) ||
+- die_die ("err", "could not bind UDP server port: $!");
++ die_die ("err", "could not bind UDP SNMP port $CF{'SNMPPORT'}: $!");
+ configure_filehandle (*SNMPSERVER) ||
+ die_die ("err", "could not configure UDP SNMP port: $!");
+ }
+@@ -2771,10 +2789,25 @@
+ my ($summary, $tmnow, $buf);
+
+ $tmnow = time;
++
++ #
++ # Since this only waits while a monitor is running, alerts which are
++ # run while no monitors are active will become zombies. They'll be
++ # reaped when a monitor is next started up. This should be harmless
++ # with any normal configuration. It could be fixed by either
++ # removing this test (and so always checking for dead kids, even
++ # when no monitors are active), or by integrating non-monitors into
++ # %runningpid and testing that rather than %running here.
++ #
+ return if (keys %running == 0);
+
+ while ((my $p = waitpid (-1, &WNOHANG)) >0)
+ {
++ if (!exists $runningpid{$p}) {
++ debug (1, "reaped non-monitor, pid $p status $?\n");
++ next;
++ }
++
+ my ($group, $service) = split (/\//, $runningpid{$p});
+ my $sref = \%{$watch{$group}->{$service}};
+
+@@ -4553,14 +4586,39 @@
+ syslog ("alert", "calling $alerttype $a for" .
+ " $args{group}/$args{service} ($alert,$args{args}) $summary");
+
+- my $pid = open (ALERT, "|-");
++ #
++ # As a simple way to avoid blocking while writing data to the alert,
++ # I fork an extra kid to do the job. First I fork one kid. This
++ # one forks another which execs the alert, then it writes the data
++ # to it. The server process continues without waiting.
++ #
++ # XXX This could be done more efficiently by forking a single kid to
++ # exec the alert, then using the main select loop to write data to
++ # it.
++ #
++ my $pid = fork;
+ if (!defined $pid) {
+ syslog ('err', "could not fork: $!");
+ return undef;
+ }
+
+ #
+- # child, the actual alert
++ # first child, fork off the alert and feed data to it
++ #
++ if ($pid == 0) {
++ $pid = open (ALERT, "|-");
++ if (!defined $pid) {
++ syslog ('err', "could not fork: $!");
++ exit;
++ }
++ if ($pid != 0) {
++ print ALERT $args{"output"};
++ exit;
++ }
++ }
++
++ #
++ # second child, the actual alert
+ #
+ if ($pid == 0) {
+ #
+@@ -4623,13 +4681,6 @@
+ }
+
+ #
+- # this will block if the alert is sucking gas
+- #
+- print ALERT $args{"output"};
+- close (ALERT);
+- waitpid $pid, 0;
+-
+- #
+ # test alerts don't count
+ #
+ return (1) if ($args{"flags"} & $FL_TEST);
+--- mon-0.99.2.orig/alert.d/mail.alert
++++ mon-0.99.2/alert.d/mail.alert
+@@ -43,7 +43,7 @@
+ $t = localtime($opt_t);
+ ($wday,$mon,$day,$tm) = split (/\s+/, $t);
+
+-open (MAIL, "| /usr/lib/sendmail -oi -t") ||
++open (MAIL, "| /usr/sbin/sendmail -oi -t") ||
+ die "could not open pipe to mail: $!\n";
+ print MAIL <<EOF;
+ To: $mailaddrs
+--- mon-0.99.2.orig/alert.d/netpage.alert
++++ mon-0.99.2/alert.d/netpage.alert
+@@ -41,7 +41,7 @@
+
+ $ALERT = $opt_u ? "UPALERT" : "ALERT";
+
+-open (MAIL, "| /usr/lib/sendmail -oi -t") ||
++open (MAIL, "| /usr/sbin/sendmail -oi -t") ||
+ die "could not open pipe to mail: $!\n";
+
+ print MAIL <<EOF;
+--- mon-0.99.2.orig/clients/monshow
++++ mon-0.99.2/clients/monshow
+@@ -1160,7 +1160,12 @@
+ @ @<<<<<<<<<<<<<< @<<<<<<<<<<< @<<<<<<<<< @<<<<<<< @<<<<<<<<< @<<< @
+ EOF
+ chomp $fmt;
+- $fmt .= "<" x length($SUMMARY) . "\n";
++ my $summary_len = do {
++ # avoid taint
++ $SUMMARY =~ /^(.*)/s or die;
++ length $1
++ };
++ $fmt .= "<" x $summary_len . "\n";
+ $fmt .= <<'EOF';
+ $DEP, $GROUP, $SERVICE, $STATUS, $TIME, $NEXT, $ALERTS, $SUMMARY
+ .
+--- mon-0.99.2.orig/clients/skymon/skymon
++++ mon-0.99.2/clients/skymon/skymon
+@@ -233,7 +233,7 @@
+ # &load_address() || die "could not load address\n";
+ # print "$BUF";
+
+- open (MAIL, "| /usr/lib/sendmail -oi -t") ||
++ open (MAIL, "| /usr/sbin/sendmail -oi -t") ||
+ die "could not open pipe to mail: $!\n";
+ print MAIL <<EOF;
+ To: $ADDR
+--- mon-0.99.2.orig/debian/debug.cf
++++ mon-0.99.2/debian/debug.cf
+@@ -0,0 +1,75 @@
++# $Id$
++
++# I use this configuration when debugging. It's still too rudimentary,
++# though, I haven't done any serious work on the daemon since I added
++# it. Still, it improves incrementally.
++
++dtlogging = yes
++randstart = 15s
++serverport = 23000
++snmp = yes
++snmpport = 23001
++syslog_facility = local0
++trapport = 23000
++
++hostgroup router
++ 208.33.154.1
++
++hostgroup http
++ jones
++ quake.argon.org
++ www.debian.org
++
++watch jones
++ service long-status
++ interval 30s
++ monitor long-status.monitor
++ service debug-listen
++ interval 30s
++ failure_interval 10s
++ monitor tcp.monitor -p 23456
++ service freespace
++ interval 1m
++ monitor freespace.monitor /:5000 /var:30000 ;;
++ service nntp
++ interval 60m
++ monitor nntp.monitor
++ service rpc
++ interval 1m
++ monitor rpc.monitor -r status -r mountd -r nfs
++
++watch non-host
++ service dns
++ interval 1m
++ monitor dns.monitor -zone argon.org -master janus hogan jones-ext ;;
++ service reboot
++ interval 1m
++ monitor reboot.monitor localhost ;;
++
++watch http
++ service ping
++ interval 10m
++ monitor fping.monitor
++ period wd {Sun-Sat}
++ alert mail.alert roderick
++ alertevery 4h
++ service http
++ depend http:ping
++ interval 10m
++ monitor http.monitor
++ period wd {Sun-Sat}
++ alert mail.alert roderick
++ alertevery 4h
++
++watch router
++ service ping
++ interval 2m
++ monitor fping.monitor
++ period wd {Sun-Sat}
++ alert mail.alert roderick
++ alertevery 4h
++
++watch bugs.debian.org
++ service ldap
++ interval 1m
++ monitor ldap.monitor --basedn 'ou=Bugs,o=Debian Project,c=US' --port 35567 --attribute bugid --filter '(bugid=1399)' --value 1399
+--- mon-0.99.2.orig/mon.d/dialin.monitor.wrap.c
++++ mon-0.99.2/mon.d/dialin.monitor.wrap.c
+@@ -11,6 +11,7 @@
+ *
+ */
+
++#include <stdio.h>
+ #include <unistd.h>
+
+ #ifndef REAL_DIALIN_MONITOR
+@@ -20,10 +21,11 @@
+ int
+ main (int argc, char *argv[])
+ {
+- char *real_img = REAL_DIALIN_MONITOR;
++ char *me = argv[0];
+
+- argv[0] = real_img;
+-
+- /* exec */
+- execv (real_img, argv);
++ argv[0] = REAL_DIALIN_MONITOR;
++ execv (argv[0], argv);
++ fprintf (stderr, "%s: can't exec ", me);
++ perror (argv[0]);
++ return 1;
+ }
+--- mon-0.99.2.orig/mon.d/file_change.monitor
++++ mon-0.99.2/mon.d/file_change.monitor
+@@ -1,4 +1,4 @@
+-#!/usr/local/bin/perl
++#!/usr/bin/perl
+ #
+ # mon monitor to watch for file changes
+ #
+--- mon-0.99.2.orig/mon.d/fping.monitor
++++ mon-0.99.2/mon.d/fping.monitor
+@@ -61,8 +61,13 @@
+
+ my @unreachable;
+ my @alive;
+-my @addr_not_found;
+ my @slow;
++my @other_prob; # details for other per-host problems
++my @error; # other errors which I'll give non-zero exit for
++my @icmp; # ICMP messages output by fping
++my %addr_unknown;
++
++my %want_host = map { $_ => 1 } @ARGV; # hosts fping hasn't output yet
+
+ while (<IN>)
+ {
+@@ -70,10 +75,15 @@
+ if (/^(\S+).*unreachable/)
+ {
+ push (@unreachable, $1);
++ delete $want_host{$1}
++ or push @error, "unreachable host `$1' wasn't asked for";
+ }
+
+ elsif (/^(\S+) is alive \((\S+)/)
+ {
++ delete $want_host{$1}
++ or push @error, "reachable host `$1' wasn't asked for";
++
+ if ($opt{"s"} && $2 > $opt{"s"})
+ {
+ push (@slow, [$1, $2]);
+@@ -87,74 +97,113 @@
+
+ elsif (/^(\S+)\s+address\s+not\s+found/)
+ {
+- push @addr_not_found, $1;
++ $addr_unknown{$1} = 1;
++ push @other_prob, "$1 address not found";
+ push @unreachable, $1;
++ delete $want_host{$1}
++ or push @error, "unknown host `$1' wasn't asked for";
++ }
++
++ # ICMP Host Unreachable from 1.2.3.4 for ICMP Echo sent to 2.4.6.8
++ # (among others)
++
++ elsif (/^ICMP (.*) for ICMP Echo sent to (\S+)/)
++ {
++ push @icmp, $_;
+ }
+
+ else
+ {
+- print STDERR "unidentified output from fping: [$_]\n";
++ push @error, "unidentified output from fping: [$_]";
+ }
+ }
+
++for my $host (keys %want_host) {
++ push @other_prob, "$host not listed in fping's output";
++ push @unreachable, $host;
++}
++
+ close (IN);
+
+ $END_TIME = time;
+
+ my $retval = $? >> 8;
+
+-if ($retval == 3)
++if ($retval < 3)
+ {
+- print "fping: invalid cmdline arguments [$CMD @ARGV]\n";
+- exit 1;
++ # do nothing
+ }
+
+-elsif ($retval == 4)
++elsif ($retval == 3)
+ {
+- print "fping: system call failure\n";
+- exit 1;
++ push @error, "fping: invalid cmdline arguments [$CMD @ARGV]";
+ }
+
+-elsif ($retval == 1 || $retval == 2 || @slow != 0)
++elsif ($retval == 4)
+ {
+- print join (" ", sort (@unreachable, map { $_->[0] } @slow)), "\n\n";
++ push @error, "fping: system call failure";
+ }
+
+-elsif ($retval == 0)
++else
+ {
+- print "\n";
++ push @error, "unknown return code ($retval) from fping";
+ }
+
+-else
+-{
+- print "unknown return code ($retval) from fping\n";
++if (@error) {
++ print "unusual errors\n";
++}
++else {
++ my @fail = sort @unreachable, map { $_->[0] } @slow;
++ # This line is intentionally blank if there are no failures.
++ print "@fail\n";
+ }
+
++print "\n";
+ print "start time: " . localtime ($START_TIME) . "\n";
+ print "end time : " . localtime ($END_TIME) . "\n";
+-print "duration : " . ($END_TIME - $START_TIME) . " seconds\n\n";
++print "duration : " . ($END_TIME - $START_TIME) . " seconds\n";
++
++if (@error != 0)
++{
++ print <<EOF;
++
++------------------------------------------------------------------------------
++unusual errors
++------------------------------------------------------------------------------
++EOF
++ print join ("\n", @error), "\n";
++}
+
+ if (@unreachable != 0)
+ {
+ print <<EOF;
++
+ ------------------------------------------------------------------------------
+ unreachable hosts
+ ------------------------------------------------------------------------------
+ EOF
+- print join ("\n", @unreachable), "\n\n";
++ print join ("\n", @unreachable), "\n";
+
+- if (@addr_not_found != 0)
+- {
+- print "address not found for @addr_not_found\n";
+- }
++ print "\nother problems:\n", join "\n", @other_prob, ''
++ if @other_prob;
++}
+
+- print "\n";
++if (@icmp != 0)
++{
++ print <<EOF;
++
++------------------------------------------------------------------------------
++ICMP messages
++------------------------------------------------------------------------------
++EOF
++ print join "\n", @icmp, '';
+ }
+
+
+ if (@slow != 0)
+ {
+ print <<EOF;
++
+ ------------------------------------------------------------------------------
+ slow hosts (response time which exceeds $opt{s}ms)
+ ------------------------------------------------------------------------------
+@@ -171,6 +220,7 @@
+ if (@alive != 0)
+ {
+ print <<EOF;
++
+ ------------------------------------------------------------------------------
+ reachable hosts rtt
+ ------------------------------------------------------------------------------
+@@ -187,25 +237,28 @@
+ #
+ if ($opt{"T"} && @unreachable)
+ {
++ my $header_output = 0;
+ foreach my $host (@unreachable)
+ {
++ next if $addr_unknown{$host};
++ print $header_output++ ? "\n" : <<EOF;
++
++------------------------------------------------------------------------------
++traceroute to unreachable hosts
++------------------------------------------------------------------------------
++EOF
+ system ("traceroute -w 3 $host 2>&1");
+ }
+-
+- print "\n";
+ }
+
++exit 1 if @error;
++
+ #
+ # fail only if all hosts do not respond
+ #
+ if ($opt{"a"})
+ {
+- if (@unreachable == @ARGV)
+- {
+- exit 1;
+- }
+-
+- exit 0;
++ exit(@alive ? 0 : 1);
+ }
+
+ exit 1 if (@slow != 0);
+--- mon-0.99.2.orig/mon.d/up_rtt.monitor
++++ mon-0.99.2/mon.d/up_rtt.monitor
+@@ -1,4 +1,4 @@
+-#!/usr/local/bin/perl
++#!/usr/bin/perl
+ #
+ # mon monitor to check for circuit up and measure RTT
+ #
projects / packages / mon.git / commitdiff