I've recently started putting together datastore clusters and implementing storage DRS. The error above does not lead to a obvious conclusion.
In the case of a "Recent Task" named "Execute Storage vMotion for Storage DRS" and error "A general system error occurred: Failed to start disk migrations." appears to be caused by vmdk's associated w/ a singular VM being named the same.
In my case, my vCOPs UI VM has 3 disks, two on one datastore, and the third w/ the config files. Storage DRS acting upon consolidation of vmdk's and maintaining 80% free space has determined that it needs to move the disks to where the vmx and 1 of the vmdk's resides. This is where the conundrum occurs and the above useless error presents itself. Below are screenshots examples:
Break the Storage DRS rule "if you have the space to do so" move all vmdk's and vmx files to a completely different datastore. This will consolidate all files to meet the consolidate vmdk rule, then storage DRS should pick up and move all files to a datastore w/ enough space bringing you back into full compliance. All without any downtime. :)