Workflow and ServiceDesk Community

This is just an observation, and may needed to be tested on more servers

We were running the ServiceDesk v7.2 and were noticing that the RAM usage was pretty high. Server had 16GB on it and almost always was running close to 15GB, CPU wasn't very busy 70~80%, but there was high disk activity, page file usage and other indicators that we decided to add more RAM.

Well, we added another 16GB (just because it was cheap), and  right off the bat, we noticed the server started to max out around 12GB, (same user load and all other variables). It was kind of surprising and lesson learned at the same time that we did not need another 16GB, could have been fine with just an additional 4GB.

Anyway this is what we observed and I wanted to share with other SD admin that might be facing similar situation.

Just curious, how was that memory usage distributed between application pool(s) and Symantec Workflow service?

Is the machine a VM or physical?

I assume that SQL is off-box, is that correct?

How many users might be hitting it at the same time?

How many CPUs/Cores does the machine have?

I can say that we run SD in VMware with an off-box physical SQL cluster. The server has 16 GB allocated to it and 8 CPUs. We have a small group of users (maybe 10 at a time) using it but never go above 30% CPU and always run around 5 GB of RAM in use.

toomas, no configuration is/was done to distribute between app pools and WF service. I suppose the OS gets to manage that...

jpellet2, This is a physical HP server, DL360, dual processors E5540 @2.53GHz (eight core each), 32GB ram, 64bit Win 2008 R2 Enterprise.

SQL is off box on exact same hardware, except 27GB dedicated to SQL, and 5GB to OS.

Max user count is close to 50, with total user count around four thousands.

And yes I was surprised to see the RAM calm down so much... and that made me share the info.