Wget HEAD request?

It's not wget, but you can do that quite easily by using curl.

curl -I http://www.superuser.com/

Produces this output:

HTTP/1.1 301 Moved Permanently                        
Content-Length: 144                       
Content-Type: text/html; charset=UTF-8     
Location: http://superuser.com/
Date: Sat, 09 Oct 2010 19:11:50 GMT


wget -S --spider www.example.com

You can also pass -O /dev/null to prevent wget from writing HTTP response to a file.

There isn't any need for curl.

With Wget, adding --spider implies that you want to send a HEAD request (as opposed to GET or POST).

This is a great minimalistic way of checking if a URL responds or not. You can for example use this in scripted checks, and the HEAD operation will make sure you do not put any load on neither the network nor the target webserver.

Bonus information: If Wget gets an HTTP error 500 from the server when it performs the HEAD it will then move on to perform a GET against the same URL. I don't know the reasoning for this design. This is the reason why you may see both a HEAD and a GET request being performed against the server. If nothing is wrong then only a HEAD request is performed. You can disable this functionality with the --tries option to limit Wget to only one attempt.

All in all, I recommend this for testing if an URL is responding:

# This works in Bash and derivatives
wget_output=$(wget --spider --tries 1 $URL  2>&1)

if [ $wget_exit_code -ne 0 ]; then
    # Something went wrong
    echo "$URL is not responding"
    echo "Output from wget: "
    echo "$wget_output"
    echo "Check succeeded: $URL is responding"