0

I am trying to setup a load balancer for few webservers. I am using centos 5.6 with apache http server (Server version: Apache/2.2.3). I am having a client generator which generates around 200 threads and try to browse through the webpage via the load balancer. I used the apache mod_proxy_balancer for this purpose. This is my setup :

<IfModule mod_proxy_balancer.c> <Proxy balancer://xx.xx.xx.xx> BalancerMember yy.yy.yy.yy loadfactor=1 BalancerMember zz.zz.zz.zz loadfactor=1 </Proxy> <Location /> ProxyPass balancer://xx.xx.xx.xx/ </Location> </IfModule> 

After the start, in few minutes I start seeing 503 errors on the client side, The apache error logs read :

[Mon Aug 22 10:19:09 2011] [error] (110)Connection timed out: proxy: HTTP: attempt to connect to yy.yy.yy.yy:80 failed [Mon Aug 22 10:19:09 2011] [error] ap_proxy_connect_backend disabling worker for (yy.yy.yy.yy) [Mon Aug 22 10:19:10 2011] [error] proxy: BALANCER: (balancer://xx.xx.xx.xx). All workers are in error state 

I changed the timeout period in the httpd.conf to 1200 and im still experiencing the same issue. I can understand that there is a timeout and it marks the ip as in error state and retries to connect after a specific time (60 seconds as per the document). It connects and continues to function well for few more minutes and this process goes on in a loop. I want to know how to fix this issue and understand what causes the timeout. If I connect the client directly to one of the backend servers , it works fine . Thus , I guess the loadbalancer causes the bottleneck. Any information to help me understand or solve this issue is greatly appreciated. Thank you !

  • Sethu

2 Answers 2

0

I am usually doing this with AJP and JBoss app server, but I have configured it something like this:

<Proxy balancer://ajpcluster> BalancerMember ajp://test.local.host:8109 BalancerMember ajp://test.local.host:8209 </Proxy> <Location /jconsole > ProxyPass balancer://ajpcluster/jconsole stickysession=JSESSIONID|jsessionid nofailover=On ProxyPassReverse balancer://ajpcluster/jconsole </Location> 
2
  • what is nofailover? Do you any errors while running the load balancer ? Commented Aug 24, 2011 at 17:58
  • @sethu nofailover is what it says it is in the documentation. Commented Jul 3, 2012 at 7:04
0

Perhaps you need whitelist the IP of the server with mod_proxy_balancer on the BalancerMember servers' firewall through which Apache is connecting out to the nodes with?

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.