Good afternoon,
We have 2 Workflows for 4/0 and 4/4
We are running 96 GB memory , 6 cores Xeon Gold 6240R 2,40 ghz
As soon as we throw in like 50000 records, our CPU workload jumps up to 100 and we’ve noticed that this interfers with other processes, basically ‘skipping’ tasks we had for other jobs, like using an escape list.
This makes it somewhat untrustworthy, anything we can do to make sure it just does all the tasks 100%? rather than decides to skip some things by itself
“We have 2 Workflows for 4/0 and 4/4” What is 4/0, 4/4?
When you say skipping task, you mean other Workflow processes are being put on hold for that one to finish?
If so, Workflow doesn’t make such decision. Each processes run simultaneously.
Now 50 000 records doesn’t seem much, unless it is XML which can increase up to 10 times the memory needed due to how Workflow handle it.
If your CPU reach and stays at 100%, I have seen weird thing happening that are impossible to reproduce in normal times.
You might want to check your workload VS your resources and maybe either increase said resources or see if some timing could be set as to when to run some jobs.
I suggest you open a ticket via our website and have a support agent review your current setup and workload.
Hi Hamelj, I’m a collegue of TBaas, maybe can give some more information.
We have a workflow with PDF input. The input is A4 double sided with about 150k pages. Throught the workflow we use the input to set it to a new output: 320x450 double sided, and 2 markers via additional content. When ever we use this workflow the CPU workload jumps to 100%. Other processes during this time will result in a error or corrupt output.
We can controle the memory through the workflow service console, but there is nothing we can do to cap the usage of the CPU, so that other processes wont go corrupt.
Is there anything we can do to make sure the CPU workload doesn’t jump to 100%. We have tried de PDF splitter: parts of 5000, parts of 1000, even 250. but the CPU still is fully utilized.
What can we do to solve this problem?
You are using Connect Template to do this?
You should look at the Connect Server engines preferences to figure out the best setup for your case.
Also when you run the job, have your task manager open and check which if the Connect engine take that much CPU.
Then from that you’ll be able to figure out which settigns need be changed or which resources to increase.
Again I reiterate that it would be better for you to open a ticket via our website and have a support agent review your current setup and workload.
I followed you instructions, it seems to be the merger and weaver engine.
So if i limit these engines in the preferences, the problem should be solved?
this seems to have some improvement, but is it normal for a weaver engine to use up to 48% of the cpu?
Unfortunately, I do not have relevant data to answer that. Maybe some of my collegue can answer this?
@Erik
@Sander
@Phil
Are you mean you want a workaround to update you with the progress for your workflow processed ?
This is my question: Is it normal for 1 weaver engine to use up to 48% of the cpu?
I don’t know enough about this to help, sorry.
In my option , the solution developed with Java and it’s normal to consume more CPU and RAM like any java application .