Many thanks for the sugestions Merlin, going to check out the HEAD
method one.

Pedro.

On Jul 4, 2009, at 8:47 PM, merlin corey wrote:

> Hello,
>
> You can change HAProxy's health check request method to HEAD. This
> would be the easiest/best solution and still highly lightweight, in my
> opinion.
>
> If for some reason you cannot do that, I created an extremely simple
> module for nginx which just always replies with a "normal" OPTIONS
> reply. You can download this module from
> http://www.rumandcoke.org/nginx/noptions.tar.gz ; extract it (I put it
> in src/http/modules) and append --add-module=src/http/modules/noptions
> (or your path to the module) then rebuild NginX. This has not been
> tested or used since 0.6.3? and may require some small changes for
> 0.7/0.8 (but it might not).
>
> Thanks,
> Merlin
>
> On Fri, Jul 3, 2009 at 9:45 AM, Pedro Mata-Mouros
> Fonseca<pedro.matamouros@sapo.pt> wrote:
>> Greetings,
>>
>> The following is an HAProxy option httpchk check request. It's
>> purpose is to
>> check a server's health by means of issuing a very lightweight
>> request to
>> it. It does this, by default, using OPTIONS (it's possible to
>> configure it
>> to perform any type of request, but OPTIONS is aparently the most
>> lightweight).
>>
>> Nginx unfortunately returns a 405 to this. It's not really a big
>> deal, but
>> is there anyway to "force" Nginx to return a 200 with an empty
>> payload or as
>> small as possible? Would you still consider this a lightweight
>> check to
>> perform every 2 or so seconds, on a few hundred reqs/sec website?
>>
>> 127.0.0.1 - - [03/Jul/2009:17:34:21 +0100] "OPTIONS / HTTP/1.0" 405
>> 166 "-"
>> "-"
>>
>> Thanks for your time.
>>
>> Pedro.
>>
>>
>
>