When a server goes below the "allowed" minimum disk space specified on the KVS admin panel, tasks still try to push content to these servers, which causes the task to be skipped. Wouldn't it make more sense for this check to be done before attempting to push content to these servers to prevent tasks for being stuck?
The main issue here is that the tasks are stuck, and need to be repossessed manually, but when attempting to evenly distribute content across media servers through the "random" distribution, it can choose another server without enough disk space. This makes the disk space limit useless since running our of disk space will return the same result (tasks failing). It also causes a backlog of tasks queued until the issue is fixed.