I see this post from 2017 and wonder if there is an update:
We have a customer that just set up a new DCC server and that is the only service running on it. It's running DCC 10.2.0 and has the scheduler and worker on the same machine.
When a process hits the timeout, the task continues working but the scheduler assigns a new task. The result is that we blow past the limit of 2 simultaneous OCR jobs and max out the CPU after running jobs for about 20 minutes. At this point, since no process gets enough resources, all of them begin to fail or time out and we get at least 5 concurrent OCR processes which all fail.
This is a specific problem for one user right now, but I feel like we run into this problem routinely with DCC. The service logs in as local system so it should have no issue terminating processes, can't think of any other set up steps we may have missed.
Does anyone have solutions or workarounds for this issue?