xref: /freebsd/contrib/ntp/scripts/monitoring/README (revision c6ec7d31830ab1c80edae95ad5e4b9dba10c47ac)
1This directory contains support for monitoring the local clock of xntp daemons.
2
3WARNING: The scripts and routines contained in this directory are beta
4	 release!  Do not depend on their correct operation. They are,
5	 however, in regular use at University of Erlangen-Nuernberg.
6	 No severe problems are known for this code.
7
8!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
9PLEASE THINK TWICE BEFORE STARTING MONITORING REMOTE XNTP DEAMONS !!!!
10MONITORING MAY INCREASE THE LOAD OF THE DEAMON MONITORED AND MAY
11INCREASE THE NETWORK LOAD SIGNIFICANTLY
12!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
13
14
15Files are:
16
17README:
18	This file
19
20ntptrap:
21	perl script to log ntp mode 6 trap messages.
22
23	It sends a set_trap request to each server given and dumps the
24	trap messages received. It handles refresh of set_trap.
25	Currently it handles only NTP V2, however the NTP V3 servers
26	also accept v2 requests. It will not interpret v3 system and
27	peer stati correctly.
28
29	usage:
30	  ntptrap [-n] [-p <port>] [-l <debug-output>] servers...
31
32	-n:		do not send set_trap requests
33
34	port:		port to listen for responses
35			useful if you have a configured trap
36
37	debug-output:	file to write trace output to (for debugging)
38
39	This script convinced me that ntp trap messages are only of
40	little use.
41
42ntploopstat:
43	perl script to gather loop info statistics from xntpd via mode 7
44	LOOP_INFO requests.
45
46	This script collects data to allow monitoring of remote xntp servers
47	where it is not possible to directly access the loopstats file
48	produced by xntpd itself. Of course, it can be used to sample
49	a local server if it is not configured to produce a loopstats file.
50
51	Please note, this program poses a high load on the server as
52	a communication takes place every delay seconds ! USE WITH CARE !
53
54	usage:
55	  ntploopstat [-d<delay>] [-t<timeout>] [-l <logfile>] [-v] [ntpserver]
56
57	delay:		number of seconds to wait between samples
58			default: 60 seconds
59	timeout:	number of seconds to wait for reply
60			default 12 seconds
61	logfile:	file to log samples to
62			default: loopstats:<ntpserver>:
63			(note the trailing colon)
64			This name actually is a prefix.
65			The file name is dynamically derived by appending
66			the name of the month the sample belongs to.
67			Thus all samples of a month end up in the same file.
68
69	the format of the files generated is identical to the format used by
70	xntpd with the loopstats file:
71		MJD <seconds since midnight UTC> offset frequency compliance
72
73	if a timeout occurs the next sample is tried after delay/2 seconds
74
75	The script will terminate after MAX_FAIL (currently 60)
76	consecutive errors.
77
78	Errors are counted for:
79		- error on send call
80		- error on select call
81		- error on recv call
82		- short packet received
83		- bad packet
84		- error on open for logfile
85
86ntploopwatch:
87	perl script to display loop filter statistics collected by ntploopstat
88	or dumped directly by xntpd.
89
90	Gnuplot is used to produce a graphical representation of the sample
91	values, that have been preprocessed and analysed by this script.
92
93	It can either be called to produce a printout of specific data set or
94	used to continously monitor the values. Monitoring is achieved by
95	periodically reprocessing the logfiles, which are updated regularly
96	either by a running ntploopstat process or by the running xntpd.
97
98	usage:
99	  to watch statistics permanently:
100	     ntploopwatch [-v[<level>]] [-c <config-file>] [-d <working-dir>]
101
102	  to get a single print out specify also
103			  -P<printer> [-s<samples>]
104				      [-S <start-time>] [-E <end-time>]
105				      [-O <MaxOffs>] [-o <MinOffs>]
106
107	level:		level of verbosity for debugging
108	config-file:	file to read configurable settings from
109			On each iteration it is checked and reread
110			if it has been changed
111			default: loopwatch.config
112	working-dir:	specify working directory for process, affects
113			interpretation of relative file names
114
115	All other flags are only useful with printing plots, as otherwise
116	command line values would be replaced by settings from the config file.
117
118	printer:	specify printer to print plot
119			BSD print systems semantics apply; if printer
120			is omitted the name "ps" is used; plots are
121			prepared using PostScript, thus the printer
122			should best accept postscript input
123
124	For the following see also the comments in loopwatch.config.SAMPLE
125
126	samples:	use last # samples from input data
127	start-time:	ignore input samples before this date
128	end-time:	ignore input samples after this date
129			if both start-time and end-time are specified
130			a given samples value is ignored
131	MaxOffs:
132	MinOffs:	restrict value range
133
134loopwatch.config.SAMPLE:
135	sample config file for ntploopwatch
136	each configurable option is explained there
137
138lr.pl:
139	linear regression package used by ntploopwatch to compute
140	linear approximations for frequency and offset values
141	within display range
142
143timelocal.pl:
144
145	used during conversion of ISO_DATE_TIME values specified in
146	loopwatch config files to unix epoch values (seconds since
147	1970-01-01_00:00_00 UTC)
148
149	A version of this file is distributed with perl-4.x, however,
150	it has a bug related to dates crossing 1970, causing endless loops..
151	The version contained here has been fixed.
152
153ntp.pl:
154	perl support for ntp v2 mode 6 message handling
155	WARNING: This code is beta level - it triggers a memory leak;
156		 as for now it is not quite clear, wether this is caused by a
157		 bug in perl or by bad usage of perl within this script.
158
159