Networking and remote

`ping`

Send ICMP echo requests to a host and report replies and round-trip times, to test reachability and latency.

ping [OPTION...] HOST

10 of 10 examples run in Debian 13.7 on 2026-10-31 · manual: ping(8)

ping host sends a small ICMP "echo request" once a second and prints each "echo reply" with its round-trip time. On Linux it keeps going until you press Ctrl-C, then prints a summary: packets sent and received, the loss percentage, and the minimum, average, maximum and deviation of the times. -c N stops after N packets, which is what you want in scripts and usually at the prompt too.

A reply proves the host is up and the network path works in both directions. No reply proves less: many firewalls drop ICMP, so a server can be serving web pages while ignoring ping. To test a specific service, connect to its port with nc -z or curl.

The sandbox has no network beyond loopback, so these examples ping 127.0.0.1 and localhost. The output format is the same for any host; real round trips are milliseconds instead of microseconds.

Examples

Send a fixed number of pings

$ ping -c 3 127.0.0.1

explain

output

PING 127.0.0.1 (127.0.0.1) 56(84) bytes of data.
64 bytes from 127.0.0.1: icmp_seq=1 ttl=64 time=0.016 ms
64 bytes from 127.0.0.1: icmp_seq=2 ttl=64 time=0.035 ms
64 bytes from 127.0.0.1: icmp_seq=3 ttl=64 time=0.036 ms

--- 127.0.0.1 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2032ms
rtt min/avg/max/mdev = 0.016/0.029/0.036/0.009 ms

Each reply line shows the sequence number, the TTL and the round-trip time. After three packets ping prints the summary: 0% loss, and rtt min/avg/max/mdev in milliseconds.

localhost may mean IPv6

$ ping -c 2 localhost

explain

output

PING localhost (::1) 56 data bytes
64 bytes from localhost (::1): icmp_seq=1 ttl=64 time=0.014 ms
64 bytes from localhost (::1): icmp_seq=2 ttl=64 time=0.047 ms

--- localhost ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1056ms
rtt min/avg/max/mdev = 0.014/0.030/0.047/0.016 ms

localhost resolved to ::1, the IPv6 loopback address, so ping used ICMPv6. Note the header line looks different: IPv6 ping reports the payload size only.

Force IPv4 or IPv6

$ ping -4 -c 2 localhost

explain

output

PING localhost (127.0.0.1) 56(84) bytes of data.
64 bytes from localhost (127.0.0.1): icmp_seq=1 ttl=64 time=0.014 ms
64 bytes from localhost (127.0.0.1): icmp_seq=2 ttl=64 time=0.028 ms

--- localhost ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1021ms
rtt min/avg/max/mdev = 0.014/0.021/0.028/0.007 ms

-4 restricts name resolution to IPv4 addresses (and -6 to IPv6), useful when a host has both and only one path is broken.

Only the summary, sent quickly

$ ping -c 10 -i 0.2 -q 127.0.0.1

explain

output

PING 127.0.0.1 (127.0.0.1) 56(84) bytes of data.

--- 127.0.0.1 ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 1865ms
rtt min/avg/max/mdev = 0.012/0.027/0.036/0.007 ms

-q suppresses the per-packet lines and -i 0.2 sends five packets a second instead of one. Very short intervals are restricted for ordinary users; the exact limit depends on the ping version.

Stop after a time limit

$ ping -w 2 127.0.0.1

explain

output

PING 127.0.0.1 (127.0.0.1) 56(84) bytes of data.
64 bytes from 127.0.0.1: icmp_seq=1 ttl=64 time=0.017 ms
64 bytes from 127.0.0.1: icmp_seq=2 ttl=64 time=0.031 ms

--- 127.0.0.1 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1007ms
rtt min/avg/max/mdev = 0.017/0.024/0.031/0.007 ms

-w 2 is a deadline: ping stops after two seconds however many packets that took, here two. Combine -c and -w to say "N replies or give up".

Send larger packets

$ ping -c 2 -s 1400 127.0.0.1

explain

output

PING 127.0.0.1 (127.0.0.1) 1400(1428) bytes of data.
1408 bytes from 127.0.0.1: icmp_seq=1 ttl=64 time=0.018 ms
1408 bytes from 127.0.0.1: icmp_seq=2 ttl=64 time=0.040 ms

--- 127.0.0.1 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1058ms
rtt min/avg/max/mdev = 0.018/0.029/0.040/0.011 ms

-s 1400 sets the payload size; ping adds 8 bytes of ICMP header and 20 of IP header, hence 1400(1428) bytes. Large pings help find links that drop big packets (MTU problems).

Add a timestamp to each reply

$ ping -c 2 -D 127.0.0.1

explain

output

PING 127.0.0.1 (127.0.0.1) 56(84) bytes of data.
[1793478028.507401] 64 bytes from 127.0.0.1: icmp_seq=1 ttl=64 time=0.015 ms
[1793478029.542808] 64 bytes from 127.0.0.1: icmp_seq=2 ttl=64 time=0.035 ms

--- 127.0.0.1 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1035ms
rtt min/avg/max/mdev = 0.015/0.025/0.035/0.010 ms

-D prefixes each line with the Unix time in brackets, handy when logging a long ping to correlate drop-outs with other events.

Check reachability in a script

$ if ping -c 1 -W 1 127.0.0.1 > /dev/null; then echo "host is up"; else echo "host is down"; fi

explain

output

host is up

ping exits 0 when at least one reply arrived. -c 1 -W 1 sends one packet and waits at most one second for it, so the check is quick even when the host is down.

No route to the host

$ ping -c 1 -W 1 10.0.0.11

explain

outputexit 2

ping: connect: Network is unreachable

The sandbox has no route beyond loopback, so ping fails immediately with Network is unreachable and exit status 2 (an error). A host that is routable but silent gives exit status 1 instead, after the timeout.

A name that does not resolve

$ ping -c 1 web01.invalid

explain

outputexit 2

ping: web01.invalid: Temporary failure in name resolution

Resolution failed before any packet was sent, which is also exit status 2. The sandbox has no DNS server, hence Temporary failure in name resolution; with working DNS, a name that does not exist gives Name or service not known. Either way, check the name and DNS rather than the network path.

Common pitfalls

Without -c, Linux ping runs forever. In scripts, always pass -c (and -W or -w for a time limit).
No reply does not mean the host is down: firewalls often block ICMP. Test the actual service port with nc -z host port or curl.
-W (timeout per reply) and -w (deadline for the whole run) are easy to mix up.

GNU vs BSD and macOS

The examples above ran the GNU/Linux version. These are the differences that matter on macOS and the BSDs (written from their documentation, not run here).

On macOS and BSD, -W is in milliseconds for some versions and -t sets the overall timeout; Linux iputils uses -W seconds and -w deadline. Check the local man page before scripting timeouts.
macOS uses a separate ping6 command for IPv6; modern Linux ping handles both.

Options worth knowing

A selection, described in our own words. For the full list, see ping(8) or ping --help.

Option	What it does
`-c COUNT`	Stop after sending COUNT packets.
`-i SECONDS`	Wait SECONDS between packets (default 1).
`-w DEADLINE`	Stop after DEADLINE seconds regardless of packet count.
`-W TIMEOUT`	Wait at most TIMEOUT seconds for each reply.
`-q`	Quiet: print only the summary.
`-s SIZE`	Payload size in bytes (default 56).
`-t TTL`	Set the IP time-to-live.
`-4`	Use IPv4 only.
`-6`	Use IPv6 only.
`-D`	Print a Unix timestamp before each line.
`-n`	Numeric output: do not look up names for addresses.
`-I INTERFACE`	Send from this interface or source address.
`-M HINT`	Path MTU discovery: do, want, dont or probe.
`-O`	Report replies that are still outstanding when the next packet is sent.
`-a`	Audible ping: beep on each reply.
`-f`	Flood ping (root only): send as fast as replies come back.

Examples

Common pitfalls

GNU vs BSD and macOS

Options worth knowing

Related