On 10/8/08 4:06 PM, "Paul Querna" <[EMAIL PROTECTED]> wrote: > I don't think the approach you are suggesting will work. > > I would suggest duplicating the workers at startup, enable/disable in > each balancer to 'move' them. > On 10/8/08 4:06 PM, "Paul Querna" <[EMAIL PROTECTED]> wrote: > Correct. So you cannot change the assignment of workers to a balancer during > runtime. > BTW: What is the goal that rises the need for you to change the assignment > of workers to balancers during runtime? >
Hi, ok. but if I cannot change the assignment of workers to balancer during runtime, how do I duplicate the workers at startup? where should I write the code for duplication? I've tried using the httpd.conf, e.g., <Proxy balancer://webapp1> BalancerMember http://ampere:81 route=ampere loadfactor=20 BalancerMember http://coulomb:81 route=coulomb loadfactor=10 BalancerMember http://hertz:81 route=hertz loadfactor=10 ProxySet lbmethod=byrequests </Proxy> <Proxy balancer://webapp2> BalancerMember http://ampere:81 route=ampere loadfactor=20 BalancerMember http://coulomb:81 route=coulomb loadfactor=10 BalancerMember http://hertz:81 route=hertz loadfactor=10 ProxySet lbmethod=byrequests </Proxy> but when Apache starts it outputs the following warning: [Wed Oct 08 18:14:48 2008] [warn] worker http://ampere:81 already used by another worker [Wed Oct 08 18:14:48 2008] [warn] worker http://coulomb:81 already used by another worker [Wed Oct 08 18:14:48 2008] [warn] worker http://hertz:81 already used by another worker Is there any problem using this approach? BTW: I'm planing to support multiple services (or applications) associated to different (dynamic) group of servers. My cluster configuration needs to be dynamic because the workload for the different running applications varies according to the time. The point is to change the overall cluster configuration (increase/decrease servers on balancer X) depending the their loads. thanks very much! Vinicius
