<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#ffffff">
<br>
-----BEGIN PGP SIGNED MESSAGE-----<br>
Hash: SHA1<br>
<br>
On 12-02-23 11:28 PM, Frank Bulk wrote:<br>
<span style="white-space: pre;">> <br>
> Starting this morning two of the sites I monitor failed, <br>
> <a class="moz-txt-link-abbreviated" href="http://www.sprint.net">www.sprint.net</a> <a class="moz-txt-link-rfc2396E" href="http://www.sprint.net"><http://www.sprint.net></a> and
<a class="moz-txt-link-abbreviated" href="http://www.sprintv6.net">www.sprintv6.net</a> <br>
> <a class="moz-txt-link-rfc2396E" href="http://www.sprintv6.net"><http://www.sprintv6.net></a>. After some testing it?s
become clear to<br>
> me that check_http doesn?t currently redirect to secure
sites, even<br>
> with the use of the '-f follow' command. I guess Sprint
turned on<br>
> redirection to the secure version this morning.<br>
> <br>
> Not sure whether to call this a bug or working-as-designed,
but I?m<br>
> request the follow feature be enhanced to allow redirection
to the<br>
> secure version of host.<br>
> <br>
> I'm running check_http version 1.4.15.<br>
> <br>
> Here's a copy of my tests:<br>
> <br>
>
==============================================================================<br>
><br>
><br>
> </span><br>
nagios:/etc/nagios3#<br>
<span style="white-space: pre;">> <br>
> nagios:/etc/nagios3# /usr/lib/nagios/plugins/check_http -H
<br>
> <a class="moz-txt-link-abbreviated" href="http://www.sprint.net">www.sprint.net</a> <a class="moz-txt-link-rfc2396E" href="http://www.sprint.net"><http://www.sprint.net></a> -4<br>
> <br>
> CRITICAL - Socket timeout after 10 seconds<br>
> <br>
> nagios:/etc/nagios3#<br>
> </span><br>
<br>
This actually has to do with <a class="moz-txt-link-abbreviated" href="http://www.sprint.net">www.sprint.net</a> not honoring the<br>
"Connection: close" header and leaving the socket open. check_http<br>
currently doesn't rely on the "Content-Length:" header to
determine<br>
when the request is complete, it rather insert a "Connection:
close"<br>
header in the request and wait for the remote server to close the<br>
socket - if the remote leaves the connection open check_http times
out.<br>
<br>
So that could be a feature request - however there is so few<br>
devices/webservers working this way that I'm wondering if it's
even<br>
valid according to the RFC's - I'd have to check.<br>
<br>
Thanks<br>
<br>
- -- <br>
Thomas<br>
-----BEGIN PGP SIGNATURE-----<br>
Version: GnuPG v1.4.10 (GNU/Linux)<br>
Comment: Using GnuPG with Mozilla - <a class="moz-txt-link-freetext" href="http://enigmail.mozdev.org/">http://enigmail.mozdev.org/</a><br>
<br>
iEYEARECAAYFAk9VrooACgkQ6dZ+Kt5Bcha7CgCdEGXPcgZySOaujHWb0dIxM4MC<br>
LecAoOJgT1s7v1yywnv6smDLmZjbnQda<br>
=dmWt<br>
-----END PGP SIGNATURE-----<br>
<br>
</body>
</html>