67c0a6ae5b643100dd5f1fa22a22774e1e9818ee
1 #!/usr/bin/perl -w
3 # (c)1999 Ian Cass, Knowledge Matters Ltd.
4 # Read the GNU copyright stuff for all the legalese
5 #
6 # Check NTP time servers plugin. This plugin requires the ntpdate utility to
7 # be installed on the system, however since it's part of the ntp suite, you
8 # should already have it installed.
9 #
10 # $Id$
11 #
12 # Nothing clever done in this program - its a very simple bare basics hack to
13 # get the job done.
14 #
15 # Things to do...
16 # check @words[9] for time differences greater than +/- x secs & return a
17 # warning.
18 #
19 # (c) 1999 Mark Jewiss, Knowledge Matters Limited
20 # 22-9-1999, 12:45
21 #
22 # Modified script to accept 2 parameters or set defaults.
23 # Now issues warning or critical alert is time difference is greater than the
24 # time passed.
25 #
26 # These changes have not been tested completely due to the unavailability of a
27 # server with the incorrect time.
28 #
29 # (c) 1999 Bo Kersey, VirCIO - Managed Server Solutions <bo@vircio.com>
30 # 22-10-99, 12:17
31 #
32 # Modified the script to give useage if no parameters are input.
33 #
34 # Modified the script to check for negative as well as positive
35 # time differences.
36 #
37 # Modified the script to work with ntpdate 3-5.93e Wed Apr 14 20:23:03 EDT 1999
38 #
39 # Modified the script to work with ntpdate's that return adjust or offset...
40 #
41 #
42 # Script modified 2000 June 01 by William Pietri <william@bianca.com>
43 #
44 # Modified script to handle weird cases:
45 # o NTP server doesn't respond (e.g., has died)
46 # o Server has correct time but isn't suitable synchronization
47 # source. This happens while starting up and if contact
48 # with master has been lost.
49 #
50 # Modifed to run under Embedded Perl (sghosh@users.sf.net)
51 # - combined logic some blocks together..
52 #
53 # Added ntpdate check for stratum 16 desynch peer (James Fidell) Feb 03, 2003
54 #
55 # ntpdate - offset is in seconds
56 # changed ntpdc to ntpq - jitter/dispersion is in milliseconds
57 #
58 # Patch for for regex for stratum1 refid.
60 require 5.004;
61 use POSIX;
62 use strict;
63 use Getopt::Long;
64 use vars qw($opt_V $opt_h $opt_H $opt_w $opt_c $opt_j $opt_k $verbose $PROGNAME $def_jitter);
65 use lib utils.pm;
66 use utils qw($TIMEOUT %ERRORS &print_revision &support);
68 $PROGNAME="check_ntp";
70 sub print_help ();
71 sub print_usage ();
73 $ENV{'PATH'}='';
74 $ENV{'BASH_ENV'}='';
75 $ENV{'ENV'}='';
77 # defaults in sec
78 my $DEFAULT_OFFSET_WARN = 60; # 1 minute
79 my $DEFAULT_OFFSET_CRIT = 120; # 2 minutes
80 # default in millisec
81 my $DEFAULT_JITTER_WARN = 5000; # 5 sec
82 my $DEFAULT_JITTER_CRIT = 10000; # 10 sec
84 Getopt::Long::Configure('bundling');
85 GetOptions
86 ("V" => \$opt_V, "version" => \$opt_V,
87 "h" => \$opt_h, "help" => \$opt_h,
88 "v" => \$verbose, "verbose" => \$verbose,
89 "w=f" => \$opt_w, "warning=f" => \$opt_w, # offset|adjust warning if above this number
90 "c=f" => \$opt_c, "critical=f" => \$opt_c, # offset|adjust critical if above this number
91 "j=s" => \$opt_j, "jwarn=s" => \$opt_j, # jitter warning if above this number
92 "k=s" => \$opt_k, "jcrit=s" => \$opt_k, # jitter critical if above this number
93 "H=s" => \$opt_H, "hostname=s" => \$opt_H);
95 if ($opt_V) {
96 print_revision($PROGNAME,'$Revision$ ');
97 exit $ERRORS{'OK'};
98 }
100 if ($opt_h) {
101 print_help();
102 exit $ERRORS{'OK'};
103 }
105 # jitter test params specified
106 if (defined $opt_j || defined $opt_k ) {
107 $def_jitter = 1;
108 }
110 $opt_H = shift unless ($opt_H);
111 my $host = $1 if ($opt_H && $opt_H =~ m/^([0-9]+\.[0-9]+\.[0-9]+\.[0-9]+|[a-zA-Z][-a-zA-Z0-9]+(\.[a-zA-Z][-a-zA-Z0-9]+)*)$/);
112 unless ($host) {
113 print "No target host specified\n";
114 print_usage();
115 exit $ERRORS{'UNKNOWN'};
116 }
118 ($opt_w) || ($opt_w = $DEFAULT_OFFSET_WARN);
119 my $owarn = $1 if ($opt_w =~ /([0-9.]+)/);
121 ($opt_c) || ($opt_c = $DEFAULT_OFFSET_CRIT);
122 my $ocrit = $1 if ($opt_c =~ /([0-9.]+)/);
124 ($opt_j) || ($opt_j = $DEFAULT_JITTER_WARN);
125 my $jwarn = $1 if ($opt_j =~ /([0-9]+)/);
127 ($opt_k) || ($opt_k = $DEFAULT_JITTER_CRIT);
128 my $jcrit = $1 if ($opt_k =~ /([0-9]+)/);
130 if ($ocrit < $owarn ) {
131 print "Critical offset should be larger than warning offset\n";
132 print_usage();
133 exit $ERRORS{"UNKNOWN"};
134 }
135 if ($opt_k < $opt_j) {
136 print "Critical jitter should be larger than warning jitter\n";
137 print_usage();
138 exit $ERRORS{'UNKNOWN'};
139 }
142 my $stratum = -1;
143 my $ignoreret = 0;
144 my $answer = undef;
145 my $offset = undef;
146 my $jitter = undef;
147 my $syspeer = undef;
148 my $candidates = 0;
149 my $msg; # first line of output to print if format is invalid
151 my $state = $ERRORS{'UNKNOWN'};
152 my $ntpdate_error = $ERRORS{'UNKNOWN'};
153 my $jitter_error = $ERRORS{'UNKNOWN'};
155 # some systems don't have a proper ntpq (migrated from ntpdc)
156 my $have_ntpq = undef;
157 if ($utils::PATH_TO_NTPQ && -x $utils::PATH_TO_NTPQ ) {
158 $have_ntpq = 1;
159 }else{
160 $have_ntpq = 0;
161 }
163 # Just in case of problems, let's not hang Nagios
164 $SIG{'ALRM'} = sub {
165 print ("ERROR: No response from ntp server (alarm)\n");
166 exit $ERRORS{"UNKNOWN"};
167 };
168 alarm($TIMEOUT);
171 ###
172 ###
173 ### First, check ntpdate
174 ###
175 ###
177 if (!open (NTPDATE, "$utils::PATH_TO_NTPDATE -q $host 2>&1 |")) {
178 print "Could not open ntpdate\n";
179 exit $ERRORS{"UNKNOWN"};
180 }
182 while (<NTPDATE>) {
183 print if ($verbose);
184 $msg = $_ unless ($msg);
186 if (/stratum\s(\d+)/) {
187 $stratum = $1;
188 }
190 if (/(offset|adjust)\s+([-.\d]+)/i) {
191 $offset = $2;
193 # An offset of 0.000000 with an error is probably bogus. Actually,
194 # it's probably always bogus, but let's be paranoid here.
195 if ($offset == 0) { undef $offset;}
197 $ntpdate_error = defined ($offset) ? $ERRORS{"OK"} : $ERRORS{"CRITICAL"};
198 print "ntperr = $ntpdate_error \n" if $verbose;
200 }
202 if (/no server suitable for synchronization found/) {
203 if ($stratum == 16) {
204 $ntpdate_error = $ERRORS{"WARNING"};
205 $msg = "Desynchronized peer server found";
206 $ignoreret=1;
207 }
208 else {
209 $ntpdate_error = $ERRORS{"CRITICAL"};
210 $msg = "No suitable peer server found - ";
211 }
212 }
214 }
216 close (NTPDATE);
217 # declare an error if we also get a non-zero return code from ntpdate
218 # unless already set to critical
219 if ( $? && !$ignoreret ) {
220 print "stderr = $? : $! \n" if $verbose;
221 $ntpdate_error = $ntpdate_error == $ERRORS{"CRITICAL"} ? $ERRORS{"CRITICAL"} : $ERRORS{"UNKNOWN"} ;
222 print "ntperr = $ntpdate_error : $!\n" if $verbose;
223 }
225 ###
226 ###
227 ### Then scan xntpq/ntpq if it exists
228 ### and look in the 11th column for jitter
229 ###
230 # Field 1: Tally Code ( Space, 'x','.','-','+','#','*','o')
231 # Only match for '*' which implies sys.peer
232 # or 'o' which implies pps.peer
233 # If both exist, the last one is picked.
234 # Field 2: address of the remote peer
235 # Field 3: Refid of the clock (0.0.0.0 if unknown, WWWV/PPS/GPS if Stratum1)
236 # Field 4: stratum (0-15)
237 # Field 5: Type of the peer: local (l), unicast (u), multicast (m)
238 # broadcast (b); not sure about multicast/broadcast
239 # Field 6: last packet receive (in seconds)
240 # Field 7: polling interval
241 # Field 8: reachability resgister (octal)
242 # Field 9: delay
243 # Field 10: offset
244 # Field 11: dispersion/jitter
245 #
247 if ($have_ntpq) {
249 if ( open(NTPQ,"$utils::PATH_TO_NTPQ -np $host 2>&1 |") ) {
250 while (<NTPQ>) {
251 print $_ if ($verbose);
252 if ( /timed out/ ){
253 $have_ntpq = 0 ;
254 last ;
255 }
256 # number of candidates on <host> for sys.peer
257 if (/^(\*|\+|\#|o])/) {
258 ++$candidates;
259 print "Candiate count= $candidates\n" if ($verbose);
260 }
262 # match sys.peer or pps.peer
263 if (/^(\*|o)([-0-9.\s]+)\s+([-0-9WwVvGgPpSs.]+)\s+([-0-9.]+)\s+([lumb]+)\s+([-0-9.]+)\s+([-0-9.]+)\s+([-0-9.]+)\s+([-0-9.]+)\s+([-0-9.]+)\s+([-0-9.]+)/) {
264 $syspeer = $2;
265 $stratum = $4;
266 $jitter = $11;
267 print "match $_ \n" if $verbose;
268 if ($jitter > $jcrit) {
269 print "Jitter_crit = $11 :$jcrit\n" if ($verbose);
270 $jitter_error = $ERRORS{'CRITICAL'};
271 } elsif ($jitter > $jwarn ) {
272 print "Jitter_warn = $11 :$jwarn \n" if ($verbose);
273 $jitter_error = $ERRORS{'WARNING'};
274 } else {
275 $jitter_error = $ERRORS{'OK'};
276 }
277 }
278 }
279 close NTPQ;
280 }
281 }
284 if ($ntpdate_error != $ERRORS{'OK'}) {
285 $state = $ntpdate_error;
286 if ($ntpdate_error == $ERRORS{'WARNING'} ) {
287 $answer = $msg . "\n";
288 }
289 else {
290 $answer = $msg . "Server for ntp probably down\n";
291 }
293 if (defined($offset) && abs($offset) > $ocrit) {
294 $state = $ERRORS{'CRITICAL'};
295 $answer = "Server Error and offset $offset msec > +/- $ocrit msec\n";
296 } elsif (defined($offset) && abs($offset) > $owarn) {
297 $answer = "Server error and offset $offset msec > +/- $owarn msec\n";
298 } elsif (defined($jitter) && abs($jitter) > $jcrit) {
299 $answer = "Server error and jitter $jitter msec > +/- $jcrit msec\n";
300 } elsif (defined($jitter) && abs($jitter) > $jwarn) {
301 $answer = "Server error and jitter $jitter msec > +/- $jwarn msec\n";
302 }
304 } elsif ($have_ntpq && $jitter_error != $ERRORS{'OK'}) {
305 $state = $jitter_error;
306 $answer = "Jitter $jitter too high\n";
307 if (defined($offset) && abs($offset) > $ocrit) {
308 $state = $ERRORS{'CRITICAL'};
309 $answer = "Jitter error and offset $offset sec > +/- $ocrit sec\n";
310 } elsif (defined($offset) && abs($offset) > $owarn) {
311 $answer = "Jitter error and offset $offset sec > +/- $owarn sec\n";
312 } elsif (defined($jitter) && abs($jitter) > $jcrit) {
313 $answer = "Jitter error and jitter $jitter msec > +/- $jcrit msec\n";
314 } elsif (defined($jitter) && abs($jitter) > $jwarn) {
315 $answer = "Jitter error and jitter $jitter msec > +/- $jwarn msec\n";
316 }
318 } elsif( !$have_ntpq ) { # no errors from ntpdate and no ntpq or ntpq timed out
319 if (abs($offset) > $ocrit) {
320 $state = $ERRORS{'CRITICAL'};
321 $answer = "Offset $offset msec > +/- $ocrit sec\n";
322 } elsif (abs($offset) > $owarn) {
323 $state = $ERRORS{'WARNING'};
324 $answer = "Offset $offset msec > +/- $owarn sec\n";
325 } elsif (( abs($offset) > $owarn) && $def_jitter ) {
326 $state = $ERRORS{'WARNING'};
327 $answer = "Offset $offset msec > +/- $owarn sec, ntpq timed out\n";
328 } elsif ( $def_jitter ) {
329 $state = $ERRORS{'WARNING'};
330 $answer = "Offset $offset secs, ntpq timed out\n";
331 } else{
332 $state = $ERRORS{'OK'};
333 $answer = "Offset $offset secs \n";
334 }
338 } else { # no errors from ntpdate or ntpq
339 if (abs($offset) > $ocrit) {
340 $state = $ERRORS{'CRITICAL'};
341 $answer = "Offset $offset msec > +/- $ocrit sec, jitter $jitter msec\n";
342 } elsif (abs($jitter) > $jcrit ) {
343 $state = $ERRORS{'CRITICAL'};
344 $answer = "Jitter $jitter msec> +/- $jcrit msec, offset $offset sec \n";
345 } elsif (abs($offset) > $owarn) {
346 $state = $ERRORS{'WARNING'};
347 $answer = "Offset $offset msec > +/- $owarn sec, jitter $jitter msec\n";
348 } elsif (abs($jitter) > $jwarn ) {
349 $state = $ERRORS{'WARNING'};
350 $answer = "Jitter $jitter msec> +/- $jwarn msec, offset $offset sec \n";
352 } else {
353 $state = $ERRORS{'OK'};
354 $answer = "Offset $offset secs, jitter $jitter msec, peer is stratum $stratum\n";
355 }
357 }
359 foreach my $key (keys %ERRORS) {
360 if ($state==$ERRORS{$key}) {
361 print ("NTP $key: $answer");
362 last;
363 }
364 }
365 exit $state;
368 ####
369 #### subs
371 sub print_usage () {
372 print "Usage: $PROGNAME -H <host> [-w <warn>] [-c <crit>] [-j <warn>] [-k <crit>] [-v verbose]\n";
373 }
375 sub print_help () {
376 print_revision($PROGNAME,'$Revision$');
377 print "Copyright (c) 2003 Bo Kersey/Karl DeBisschop\n";
378 print "\n";
379 print_usage();
380 print "
381 Checks the local timestamp offset versus <host> with ntpdate
382 Checks the jitter/dispersion of clock signal between <host> and its sys.peer with ntpq\n
383 -w ( --warning)
384 Clock offset in seconds at which a warning message will be generated.\n Defaults to $DEFAULT_OFFSET_WARN.
385 -c (--critical)
386 Clock offset in seconds at which a critical message will be generated.\n Defaults to $DEFAULT_OFFSET_CRIT.
387 -j (--jwarn)
388 Clock jitter in milliseconds at which a warning message will be generated.\n Defaults to $DEFAULT_JITTER_WARN.
389 -k (--jcrit)
390 Clock jitter in milliseconds at which a warning message will be generated.\n Defaults to $DEFAULT_JITTER_CRIT.\n
392 If jitter/dispersion is specified with -j or -k and ntpq times out, then a
393 warning is returned.
394 ";
395 support();
396 }