We are using the following haproxy version on ubuntu 10.04 haproxy -v HA-Proxy version 1.4.22 2012/08/09
When trying to run snort IDS on my test server, it hangs the entire cluster of app servers (apache2) with one app server (app3) getting all the load , my question is why aren't the other 2 app servers(app1 and app2) functioning accurately, is it because my apps internally make xmlrpc calls to the other servers and since app 3 isn't down it just hangs the entire site? my configuration for this backend is backend https mode http stats enable stats auth admin:xxxx stats uri /admin?stats balance roundrobin cookie BALANCEID insert indirect nocache option http-server-close option forwardfor except 127.0.0.1 option httpchk OPTIONS /check2013.php HTTP/1.1\r\nHost:\ www server app1 xx.xx.xx.197:80 cookie A check inter 20000 maxconn 30 server app2 xx.xx.xx.190:80 cookie B check inter 20000 maxconn 30 server app3 xx.xx.xx.55:80 cookie C check inter 20000 maxconn 30 i recently added a few things to try and see if i can eliminate a bad apache server from getting the requests... i added check2013.php which has <?php echo "I'm healthy"; ?> and i also added maxconn 30, so the other requests might get queued but none of these helped so far So i might be total offtrack here in trying to resolve this issue, hence i wanted to share this with the forum...
