Load balancing problem on cluster EJB calls in Weblogic

综合编程 2018-02-16 阅读原文

Our Application has 2 Clusters running in Weblogic 12 C.

Each Cluster 4 Managed Servers. Cluster B is making EJB Call to Cluster A through Foreign JNDI . In Cluster A EJB Calls are not distributing across the managed server ? This results in overload of One Server in Cluster A. We are using round robin load balancing in Cluster Configuration.

Can you tell me any other configuration we can make in our Weblogic to make ejb calls get load balanced ?

Regards, Sakumar

It's like you need to set your Foreign JNDI URL to specify all of the hosts in the cluster. It doesn't matter if the cluster is load balanced if the client keeps calling the same host.

Instead of this, which will direct all traffic to one host:

java.naming.provider.url = t3://server1:port1

Try using:

java.naming.provider.url = t3://server1:port1,server2:port2


It sounds like the weblogic algorithm is not doing what you want, look at: https://docs.oracle.com/cd/E24329_01/web.1211/e24425/load_balancing.htm#CLUST180

Specifically - the round-robin algorithm cycles through a list of WebLogic Server instances in order
. So it's hitting the first server every time because that is the first in your list. I would recommend trying another algorithm or using an external load balancer

Hello, buddy!

责编内容by:Hello, buddy!阅读原文】。感谢您的支持!


The need for load balancing in medical imaging Medical information and imaging systems are expected to operate like a luxury automobile. You get in, turn the key, it s...
CVE-2018-2628 Weblogic反序列化POC重构详解 01 背景 — Oracle官方在修补老的Weblogic反序列化漏洞时,将已经在互联网上暴露的PoC都加入了黑名单,所以若是绕过他的黑名单限制就依旧可以触发命令执行。接下来我们简单...
Taking Zero-Downtime Load Balancing even Further Ever since we rolled out our zero-downtime HAProxy reload system a few years ago, we have been disappointed that it req...
分布式与集群的总结 概念 “分头做事”与“一堆人”的区别 分布式:有一堆人,来了一个业务,划分为多个子业务,一人一个分着做完。 集群:有一堆人,来了一个业务,不管给谁都能一个人单独做完。 分布式 ...
网易MySQL中间件的负载均衡策略及性能优化... 一、背景 随着业务的爆发式增长,电商系统中的读写压力越来越高,单节点MySQL实例压力越来越大,单纯升级服务器硬件已经无法满足生产环境的需要。解决读请求压力,需要支持从库扩展;解决写请求压力,对数据分片增加多个节点,降低单节点MySQL...