Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

What IP address is OK to ping -f for a prolonged time (the whole day)? "" or something else?

The main concern is the admins there that may decide that it is some DoS and block/complain...

Maybe there is special dedicated public test server somewhere that sends ICMP replies (including fast enough), runs TCP and UDP echo servers and other tools?..

share|improve this question
One you own and whose CPU and network usage you pay for? Ping is probably the wrong tool for monitoring connectivity outages. – RedGrittyBrick Sep 18 '12 at 14:07
(My own cheap VPS is a bit unreliable). What is the good tool to catch short transient network outages (not needing a special server or that have pulic servers)? – Vi. Sep 18 '12 at 14:10
Imho icmp can be a good tool to test network outages. Normal internet-usage like surfing only creates a problem when you have several packet-losts. – Langhard Sep 18 '12 at 14:12
My ADSL router logs shows DSL disconnections and re-connections. Server syslog entries may show symptoms of connection failures for various services. – RedGrittyBrick Sep 18 '12 at 14:13
@Riscie: man ping: "This program is intended for use in network testing, measurement and management. Because of the load it can impose on the network, it is unwise to use ping during normal operations or from automated scripts." – RedGrittyBrick Sep 18 '12 at 14:15
up vote 6 down vote accepted

EDIT: **As pointed out by RedGrittyBrick and Hennes: ping -f is not a good idea on any public ip!

ping -t instead should imho not be a problem as this does not generate much network load.

I always use Google's DNS-Server-Adresses:


and never encountered any problems with packet-losts etc.

share|improve this answer
I use this and for pings, but I'm not sure about more prolonged and/or high-bandwidth tests. Is it really OK? – Vi. Sep 18 '12 at 14:06
I had one test running for 5 days with no problems on Note sure what to say about high-bandwidth tests but icmp shouldn't be a problem at all. – Langhard Sep 18 '12 at 14:08
I always use Google DNS as well. – Dave Sep 18 '12 at 14:14
Should I consider, for example, ping -f -i 0.2 low-bandwidth enough for this? – Vi. Sep 18 '12 at 14:15
Where -t stands for timespan (eg ping -t 120 would ping for 120 seconds. Leaving the time off implies a continuous ping) – Dave Sep 18 '12 at 14:55

While I want to stress that leeching on someone else's service like this in a permanent setup is against my personal ethics, I will supply this pointer.

ICMP is often filtered or limited in rate of requests per source address, but HTTP services generally are not. You could use nping from nmap which (by default) uses TCP handshakes on port 80 rather than ICMP echoes to detect connectivity. As a bonus it works without being root on UNIX systems. The man page describes many, many options for it.

You'll have to either find a static address to a web server with good uptime (I have no idea how to tell if Google changes their front-end addresses) or rely on both a dns server AND a webserver. That might be an issue for your setup.

share|improve this answer
Replying to HTTP request is even more load than just replying the ICMP. And [semi-]flooded HTTP requests seem even more DoS'sy that flooded ICMPs. (Semi-flooded here means 4-40 requests per second). – Vi. Sep 18 '12 at 15:38
nping opens a TCP connection, but doesn't issue an HTTP request. – jdizzle Apr 30 '14 at 1:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .