[Nagiosplug-help] Change in check_http behavior, and how to disable?

[Chris Knight]

For over a year I've had a fairly simple http check in place, and one
day a few weeks back it suddenly stopped working after an update of my
nagios plugins.  The check is fairly simple:

/opt/local/libexec/nagios/check_http -H download.gamevideos.com -u /favicon.ico

The check fails, but using the URL
http://download.gamevideos.com/favicon.ico in wget, or a web browser,
works fine.

I tracked down the issue to a change in the Host header that
check_http is sending.  It appears that in a recent update to
check_http started adding the default port (80) to the Host header.

So, where wget issues:

wget http://download.gamevideos.com/favicon.ico

  GET /favicon.ico HTTP/1.0
  User-Agent: Wget/1.11.4
  Accept: */*
  Host: download.gamevideos.com
  Connection: Keep-Alive

check_http sends this:

/opt/local/libexec/nagios/check_http -H download.gamevideos.com -u /favicon.ico

  GET /favicon.ico HTTP/1.0
  User-Agent: check_http/v1991 (nagios-plugins 1.4.12)
  Connection: close
  Host: download.gamevideos.com:80

While RFC 2616 does indicate that adding :PORT to the Host header is
valid, this can be a problem when the website in question is behind a
content switch that is doing exact matches on the Host header.  In my
case, I wound up adding an additional set of rules to match the
host:port host header, but it sure would have been nice to have a
command line flag I could have used on check_http to disable this new
behavior.

Just wondering if this has bitten anyone else, and thought I would
share in case this was an active problem for anyone.

-Chris