|
|
| |
MRTG-PING-PROBE(1) |
User Contributed Perl Documentation |
MRTG-PING-PROBE(1) |
mrtg-ping-probe - a round trip time and packet loss probe for MRTG
mrtg-ping-probe [ -hsvV ] [ -d deadtime ] [
-k count ] [ -l length ] [ -o
ping_options ] [ -p
[factor*]item/[factor*]item ] [ -r
[rsh:][user@]host[:osname] ] [ -t timeout ] host
mrtg-ping-probe pings the given host host and prints on stdout two
lines extracted from the ping output. The default is to print the maximum, and
the minimum round trip time.
It is meant to be called by the Multi Router Traffic Grapher
(MRTG).
- -h
- print help on stdout and exit.
- -v
- Be more verbose.
- -V
- Print version number on stderr and exit.
- -d deadtime
- Specifies the value we return for round trip times in case we assume that
the target is down. The default is zero. We assume that the target is
unreachable, if we cannot find the ping summary or if the ping program was
aborted because of a set timeout.
For WAN connections that usually have round trip times of 10ms
and higher, ranges of zero round trip time are highly visible. In a LAN
environment, you might set it to a high value, e.g. 999, which however
might change the scale of the graphs in such a way that you hardly see
the regular round trip times. You might use mrtg-misc-probe's pong
option to generate a graph that shows reachability of targets,
instead.
- -k count
- Specifies the number of of ping packets to be sent. The default is to send
10 ping packets.
- -l length
- Use length as the length of the data portion of the ICMP ECHO
request packet. The default length is 56 data bytes.
- -o ping_options
- Pass ping_options to the ping program. You can use this generic
option to e.g. pass an option to ping to suppress displaying addresses as
host names. This helps to prevent the ping to fail because it cannot map
hostnames to IP addresses and vice versa. To pass several arguments,
enclose the options in quotes. Check the documentation of your ping
program for possible options.
- -p
[factor*]item/[factor*]item
- Pick the values you want mrtg-ping-probe to return. Allowed values for
item are: min, max, avg, loss, or an
integer. Each item can be preceded by a integer factor used to
multiply the value returned by the ping program. The default pick-list is
min/max.
To display ping times in microseconds instead of milliseconds,
use: -p 1000*max/1000*min.
- -r [rsh:][user@]host[:osname]
- Not Yet Implemented
run ping on remote host host, as user user (or
as local user, if no user is given). Uses rsh -n to start program
on remote host, unless you provide a different program name. If the
remote host has a different system type than the local host (if the
osname is different) you have to say so.
This option can be used if you run mrtg on a host that cannot
ping to the final target, and you cannot install mrtg and/or perl on the
intermediate host used to ping the final target.
- -s
- Silent mode. Do not generate error messages if there is no response from
the ping program or if it ran into the timeout. Usually cron will mail you
these error messages, which might be helpful to debug problems.
- -t timeout
- Abort the external ping program after timeout seconds. A
timeout value of zero (the default) means, we do not abort the
external ping program.
If mrtg-ping-probe seems to hang forever, check your ping
program, it might be a version that wants to receive the given
number of ECHO_RESPONSE packets instead of just sending them. If your
target is unreachable, these pings ping forever.
You want to choose timeout as short as possible to
leave mrtg enough time for all your other targets, but long enough so
you do not abort pings (too often). You might use (count * worst
case round trip time) as a starting point. (Or install a ping program
that is not broken ;-)
If your perl installation does not implement the built-in
alarm() function, the timeout option will be ignored. You will
get a warning about this only in verbose mode (option -v). I have
not found a perl installation on Windows that implements the
alarm() built-in function on Win32. So basically on Windows the
timeout option is not working.
The program exits with an exit value 0, if it believes it was successful.
- mrtg-ping-probe ricochet
- Retrieves the maximum and minimum round trip time to the host
ricochet, using the default length and count.
- mrtg-ping-probe -p max/avg ricochet
- Retrieves the maximum and average round trip time to the host
ricochet, using the default length and count.
- mrtg-ping-probe -p '1000*max/1000*avg' ricochet
- Retrieves the maximum and average round trip time to the host
ricochet multiplied by a factor of 1000, using the default length
and count.
- mrtg-ping-probe -k 17 -l 1000 192.168.192.42
- Retrieves the maximum and minimum round trip time to the host
192.168.192.42, using 17 1000 data bytes pings.
- mrtg-ping-probe -o -n ricochet.pwo.de
- Suppress displaying addresses as host names on Solaris 2 (to protect from
DNS problems causing the ping to fail) by passing option -n to the
ping program.
Note that in this example `-n' is not an option for
mrtg-ping-probe, but gets passed to the ping program.
- mrtg-ping-probe -o '-n -I 3' ricochet.pwo.de
- Pass several options -n -I 3 to the ping program.
- mrtg-ping-probe -p loss/loss ricochet.pwo.de
- Monitors the packet loss for the link to host ricochet.pwo.de.
mrtg-ping-probe uses an external ping program, like
/usr/sbin/ping.
mrtg(1), mrtg-ping-cfg, ping(1), mrtg.cfg-ping,
mrtg-misc-probe(1)
http://www.mrtg.org/
http://pwo.de/projects/mrtg/
- FATAL: Not yet configured for osname
- Currently mrtg-ping-probe depends on an external ping program,
which every operating systems hides in another place. Also different
programs require different arguments. We have a configuration table
listing the ping program for each operating systems. You have to figure
out how to call which program on your platform, and add to the information
to the table. Please contribute back any additions, so I can include them
in the next version.
- ERROR: ignoring superfluous arguments
- More than one argument was given. mrtg-ping-probe will ignore all
but the first argument. The first argument is taken as a hostname or IP
address of an host and mrtg-ping-probe will try to ping it.
- FATAL: ping what?
- No argument was given. mrtg-ping-probe terminates, as there is
nothing to ping.
- FATAL: option option requires numeric argument.
- The argument for option option was not an integer number.
- FATAL: Can't open ping: some reason
- mrtg-ping-probe was not able to execute the external ping program.
Check the pathname and permissions of the external ping program. some
reason might give some useful hints.
- ERROR: external ping hit timeout timeout, assuming target
host is unreachable
- We ran into a timeout pinging the target host host. You might have
to increase the timeout value (Option -t) if this happens when the
target is up and the round trip time just happens to be longer than usual.
The captured output of the ping program is printed and will
(hopefully) give further hints why this problem occurred.
This message is not printed if mrtg-ping-probe runs in silent
mode (Option -s).
- ERROR: Could not find ping summary for host
- mrtg-ping-probe was not able to find the ping summary. Most likely,
the host is not reachable. If your operating system changed (e.g. it was
upgraded to a new version, or a new version of the ping program was
installed), it might also be necessary to change the regular expression
that extracts the round trip times. You might want to use the perl script
check-ping-fmt (which is part of the source distribution) to test the
regular expression.
The captured output of the ping program is printed and will
(hopefully) give further hints why this problem occurred.
This message is not printed if mrtg-ping-probe runs in silent
mode (Option -s).
- ERROR: Could not find packet loss summary for host
- mrtg-ping-probe was not able to find the packet loss summary. If
your operating system changed (e.g. it was upgraded to a new version, or a
new version of the ping program was installed), it might also be necessary
to change the regular expression that extracts the packet loss. You might
want to use the perl script check-ping-fmt (which is part of the source
distribution) to test the regular expression.
The captured output of the ping program is printed and will
(hopefully) give further hints why this problem occurred.
This message is not printed if mrtg-ping-probe runs in silent
mode (Option -s).
mrtg-ping-probe currently depends on an external ping(1) program.
If the external program does not support an option, the option given to
mrtg-ping-probe will be ignored.
Under freebsd release 3.x or later, ping option -s
packet-length (mrtg-ping-probe option -l length)
is only allowed to be used when we run as root (which we should not),
therefore this option is silently removed on freebsd before we call
the external ping program.
This program has way too many options and tries to support too many different
systems.
Using this program to monitor sub-millisecond round trip times or
packet loss might be questionable.
Option -r, remote execution of the ping program, is not yet
implemented.
Copyright (c) 1997-2003 Peter W. Osel <pwo@pwo.de>. All Rights Reserved.
See the file COPYRIGHT in the distribution for the exact
terms.
Written by Peter W. Osel <pwo@pwo.de>. http://pwo.de/
Visit the GSP FreeBSD Man Page Interface. Output converted with ManDoc. |