Hi, thanks for posting.
NCrunch creates workspaces as it needs them. This means that the number of workspaces created by the tool can only be controlled indirectly by configuration.
The most significant setting in doing this is the 'Max number of processing threads' setting. When the engine has more processing threads, it will tend to have more test runners open at one time, which generally means that more workspaces will be required if projects are being constantly changed. NCrunch won't interrupt an individual test from running once it's started, and it can't tear down any workspaces being used by a test process. Thus lowering the max number of processing threads will reduce disk consumption at the expense of reduced overall performance.
Another setting that can affect the number of workspaces present at a point in time is the 'Max test runners to pool' setting. When more test runner processes are being pooled between test runs, more processes are running at any one time, and therefore more workspaces are locked.
Broadly speaking, it is not possible to reduce the number of workspaces used by NCrunch without seriously affecting the performance of the engine. If disk consumption is a problem, my advice is generally to try to increase the amount of disk space available or look at reducing the size of the projects/solution by removing any large resource files that may not be required in NCrunch's test environment. Trying to limit the disk consumption of a grid node can in most cases be rather self-defeating, as most people set up grid nodes in an effort to increase engine performance, and by limiting their disk consumption through the above settings they can greatly reduce the advantages of such a setup.
