Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

settingup load balancer server in IBM Websphere environment

I am using websphere network deployment. I have written an application that is deployed on two different nodes on the same cell. Now I want to load balance the two servers with sticky session option. Is there a way we can create load balancer server within the websphere environment? I have seen an option called On demand routing, is that equal to load balancer? Your help on this is much appreciated. Btw I am using websphere 8.5.x. thanks.

like image 582
Ram C Avatar asked Feb 09 '14 03:02

Ram C


People also ask

What is load balancer in WebSphere application server?

The load balancer is responsible for distributing incoming requests evenly to a collection of IMS Servers on the application-tier. By using a load balancer with session affinity, traffic from each client is always routed to the same IBM HTTP Server.

What type of server is an IBM WebSphere?

IBM WebSphere Application Server is a flexible, secure Java server runtime environment for enterprise applications. Deploy and manage applications and services regardless of time, location or device type.

Can we do load balancing with one server also?

It's pointless to load balance a single server. What load balancing does is decide which server to use based on some criteria (typically load). You could have a load balance installed, but it wouldn't be serving any purpose. Another major reason for multiple servers isn't load balancing, but redundancy.


1 Answers

Yes, the On Demand Router is the native load balancer for a WebSphere cell. Then you simply use the "weight" feature to control the distribution of requests.

Setting Weight - In the admin console, go to "Servers > Clusters > [your cluster type] > [your cluster name]" then under the "Additional Properties" heading click "Cluster members".

Set their "configured weight" to equal values and then ripplestart the cluster (or cycle individually). Note, "runtime weight" takes effect immediately but is reset to configured weight whenever the server is started.

Requests are allocated by a simple formula: (weight of Member X) / (Sum of all member weights)

so if server A has a weight of 20 and server B has a weight of 1, then server A should get 20 / (20+1) or 95.3% of all requests.

In your case, just set them to the same weight.

like image 98
retriever123 Avatar answered Sep 24 '22 15:09

retriever123