The default health check is an OPTIONS call. Perhaps the configuration
of the web server is different on FreeBSD and doesn't permit OPTIONS by
default.
You could either change the configuration of the web server to permit it
or change the call to a GET (though this may be a heavier request for
the server).
Search the documentation for "option httpchk <method> <uri>".
-JohnF
On 10-12-15 12:49 AM, bluethundr wrote:
Hello HaProxy list,
I have recently setup an install of haproxy on two virtual instances
in my environment. It is performing beautifully and really holding up
to it's task quite well.
To explain a bit about my setup, I have 3 CentOS 5.5 instances
hosting the site and a fourth doing the same that happens to be a
FreeBSD 8.1 server. All three CentOS instances turn up in the stats
interface with accurate statistics on their performance and uptime.
However, the FreeBSD server (which has been verified to be functioning
well) always shows as down in the stats interface.
All hosts, CentOS as well as FreeBSD share an NFS mounted web
directory so that they all serve the same exact content. You can hit
each server on the local lan through it's internal DNS and they all
function. So why would only the FreeBSD machine show as down?
Thanks!!
Tim
