Net::Ping(3perl) | Perl Programmers Reference Guide | Net::Ping(3perl) |
Net::Ping - check a remote host for reachability
use Net::Ping; $p = Net::Ping->new(); print "$host is alive.\n" if $p->ping($host); $p->close(); $p = Net::Ping->new("icmp"); $p->bind($my_addr); # Specify source interface of pings foreach $host (@host_array) { print "$host is "; print "NOT " unless $p->ping($host, 2); print "reachable.\n"; sleep(1); } $p->close(); $p = Net::Ping->new("tcp", 2); # Try connecting to the www port instead of the echo port $p->port_number(scalar(getservbyname("http", "tcp"))); while ($stop_time > time()) { print "$host not reachable ", scalar(localtime()), "\n" unless $p->ping($host); sleep(300); } undef($p); # Like tcp protocol, but with many hosts $p = Net::Ping->new("syn"); $p->port_number(getservbyname("http", "tcp")); foreach $host (@host_array) { $p->ping($host); } while (($host,$rtt,$ip) = $p->ack) { print "HOST: $host [$ip] ACKed in $rtt seconds.\n"; } # High precision syntax (requires Time::HiRes) $p = Net::Ping->new(); $p->hires(); ($ret, $duration, $ip) = $p->ping($host, 5.5); printf("$host [ip: $ip] is alive (packet return time: %.2f ms)\n", 1000 * $duration) if $ret; $p->close(); # For backward compatibility print "$host is alive.\n" if pingecho($host);
This module contains methods to test the reachability of remote hosts on a network. A ping object is first created with optional parameters, a variable number of hosts may be pinged multiple times and then the connection is closed.
You may choose one of six different protocols to use for the ping. The "tcp" protocol is the default. Note that a live remote host may still fail to be pingable by one or more of these protocols. For example, www.microsoft.com is generally alive but not "icmp" pingable.
With the "tcp" protocol the ping() method attempts to establish a connection to the remote host's echo port. If the connection is successfully established, the remote host is considered reachable. No data is actually echoed. This protocol does not require any special privileges but has higher overhead than the "udp" and "icmp" protocols.
Specifying the "udp" protocol causes the ping() method to send a udp packet to the remote host's echo port. If the echoed packet is received from the remote host and the received packet contains the same data as the packet that was sent, the remote host is considered reachable. This protocol does not require any special privileges. It should be borne in mind that, for a udp ping, a host will be reported as unreachable if it is not running the appropriate echo service. For Unix-like systems see inetd(8) for more information.
If the "icmp" protocol is specified, the ping() method sends an icmp echo message to the remote host, which is what the UNIX ping program does. If the echoed message is received from the remote host and the echoed information is correct, the remote host is considered reachable. Specifying the "icmp" protocol requires that the program be run as root or that the program be setuid to root.
If the "external" protocol is specified, the ping() method attempts to use the "Net::Ping::External" module to ping the remote host. "Net::Ping::External" interfaces with your system's default "ping" utility to perform the ping, and generally produces relatively accurate results. If "Net::Ping::External" if not installed on your system, specifying the "external" protocol will result in an error.
If the "syn" protocol is specified, the "ping" method will only send a TCP SYN packet to the remote host then immediately return. If the syn packet was sent successfully, it will return a true value, otherwise it will return false. NOTE: Unlike the other protocols, the return value does NOT determine if the remote host is alive or not since the full TCP three-way handshake may not have completed yet. The remote host is only considered reachable if it receives a TCP ACK within the timeout specified. To begin waiting for the ACK packets, use the "ack" method as explained below. Use the "syn" protocol instead the "tcp" protocol to determine reachability of multiple destinations simultaneously by sending parallel TCP SYN packets. It will not block while testing each remote host. This protocol does not require any special privileges.
"proto" specifies the protocol to use when doing a ping. The current choices are "tcp", "udp", "icmp", "icmpv6", "stream", "syn", or "external". The default is "tcp".
If a "timeout" in seconds is provided, it is used when a timeout is not given to the ping() method (below). The timeout must be greater than 0 and the default, if not specified, is 5 seconds.
If the number of data bytes ("bytes") is given, that many data bytes are included in the ping packet sent to the remote host. The number of data bytes is ignored if the protocol is "tcp". The minimum (and default) number of data bytes is 1 if the protocol is "udp" and 0 otherwise. The maximum number of data bytes that can be specified is 65535, but staying below the MTU (1472 bytes for ICMP) is recommended. Many small devices cannot deal with fragmented ICMP packets.
If "device" is given, this device is used to bind the source endpoint before sending the ping packet. I believe this only works with superuser privileges and with udp and icmp protocols at this time.
If <tos> is given, this ToS is configured into the socket.
For icmp, "ttl" can be specified to set the TTL of the outgoing packet.
Valid "family" values for IPv4:
4, v4, ip4, ipv4, AF_INET (constant)
Valid "family" values for IPv6:
6, v6, ip6, ipv6, AF_INET6 (constant)
The "host" argument implicitly specifies the family if the family argument is not given.
The "port" argument is only valid for a udp, tcp or stream ping, and will not do what you think it does. ping returns true when we get a "Connection refused"! The default is the echo port.
The "bind" argument specifies the local_addr to bind to. By specifying a bind argument you don't need the bind method.
The "gateway" argument is only valid for IPv6, and requires a IPv6 address.
The "retrans" argument the exponential backoff rate, default 1.2. It matches the $def_factor global.
The "dontfrag" argument sets the IP_DONTFRAG bit, but note that IP_DONTFRAG is not yet defined by Socket, and not available on many systems. Then it is ignored. On linux it also sets IP_MTU_DISCOVER to IP_PMTUDISC_DO but need we don't chunk oversized packets. You need to set $data_size manually.
This is enabled by default.
This affects the "udp", "tcp", and "syn" protocols.
This is disabled by default.
With argument sets the option. Without returns the option value.
With argument sets the option. Without returns the option value.
Not yet implemented.
If the protocol is set to "tcp", this method may be called any number of times, and each call to the ping() method (below) will use the most recent $local_addr. If the protocol is "icmp" or "udp", then bind() must be called at most once per object, and (if it is called at all) must be called before the first call to ping() for that object.
The bind() call can be omitted when specifying the "bind" option to new().
Without argument, it returns the currently used icmp protocol message type. By default, it returns 'echo'.
The $host argument can be omitted when specifying the "host" option to new().
When "new" had a host option, this host will be used. Without $host argument, all hosts are scanned.
Perform a stream ping. If the tcp connection isn't already open, it opens it. It then sends some data and waits for a reply. It leaves the stream open on exit.
Perform a udp echo ping. Construct a message of at least the one-byte sequence number and any additional data bytes. Send the message out and wait for a message to come back. If we get a message, make sure all of its parts match. If they do, we are done. Otherwise go back and wait for the message until we run out of time. Return the result of our efforts.
Default host: '255.255.255.255' Default port: 9
perl -MNet::Ping=wakeonlan -e'wakeonlan "e0:69:95:35:68:d2"'
There will be less network overhead (and some efficiency in your program) if you specify either the udp or the icmp protocol. The tcp protocol will generate 2.5 times or more traffic for each ping than either udp or icmp. If many hosts are pinged frequently, you may wish to implement a small wait (e.g. 25ms or more) between each ping to avoid flooding your network with packets.
The icmp and icmpv6 protocols requires that the program be run as root or that it be setuid to root. The other protocols do not require special privileges, but not all network devices implement tcp or udp echo.
Local hosts should normally respond to pings within milliseconds. However, on a very congested network it may take up to 3 seconds or longer to receive an echo packet from the remote host. If the timeout is set too low under these conditions, it will appear that the remote host is not reachable (which is almost the truth).
Reachability doesn't necessarily mean that the remote host is actually functioning beyond its ability to echo packets. tcp is slightly better at indicating the health of a system than icmp because it uses more of the networking stack to respond.
Because of a lack of anything better, this module uses its own routines to pack and unpack ICMP packets. It would be better for a separate module to be written which understands all of the different kinds of ICMP packets.
The latest source tree is available via git:
git clone https://github.com/rurban/Net-Ping.git cd Net-Ping
The tarball can be created as follows:
perl Makefile.PL ; make ; make dist
The latest Net::Ping releases are included in cperl and perl5.
For a list of known issues, visit:
<https://rt.cpan.org/NoAuth/Bugs.html?Dist=Net-Ping> and <https://github.com/rurban/Net-Ping/issues>
To report a new bug, visit:
<https://github.com/rurban/Net-Ping/issues>
Current maintainers: perl11 (for cperl, with IPv6 support and more) p5p (for perl5) Previous maintainers: bbb@cpan.org (Rob Brown) Steve Peters External protocol: colinm@cpan.org (Colin McMillen) Stream protocol: bronson@trestle.com (Scott Bronson) Wake-on-lan: 1999-2003 Clinton Wong Original pingecho(): karrer@bernina.ethz.ch (Andreas Karrer) pmarquess@bfsec.bt.co.uk (Paul Marquess) Original Net::Ping author: mose@ns.ccsn.edu (Russell Mosemann)
Copyright (c) 2017-2018, Reini Urban. All rights reserved.
Copyright (c) 2016, cPanel Inc. All rights reserved.
Copyright (c) 2012, Steve Peters. All rights reserved.
Copyright (c) 2002-2003, Rob Brown. All rights reserved.
Copyright (c) 2001, Colin McMillen. All rights reserved.
This program is free software; you may redistribute it and/or modify it under the same terms as Perl itself.
2023-11-25 | perl v5.32.1 |