Friday, 20 May 2016

impact of 80 JD Edwards users on a windows system

I’m running a fairly intensive interactive load and some batch activity – 80 concurrent users across 2 web servers and 2 enterprise servers.

Everything is running okay, I’d recommend more CPU for enterprise servers, leave the RAM at 8 for this.  You can see what the boxes are doing at the web tier and the enterprise server tier below.  JDE Ent servers are generally very light on RAM, especially when you tune your kernels appropriately.  CPU is important here, and network latency.

Enterprise Server

image

Note that there are 4 cores viewable in resource monitor.

image

 

image

Another chart that is really handy, showing the amount of disk and also network traffic that is generated from this exercise. 

image

Server 2 of my pigeon pair is also loaded up the same way.  Note that these servers have a single CPU and 8GB of RAM.

Also, this is a

Web

Note that there are 4 cores viewable in resource monitor.

image

image

No comments:

Extending JDE to generative AI