Okay, so....as mentioned above, the new attempts at jobs were just sitting there queued. VJware kindly helped me offline and figured out that there were old files in the BE temp dir on C (which if I understand correctly, were the mount points BE makes to access the snapshots). Since those were in there leftover from a previous backup attempt, the new one was not starting properly. Once those were cleared out, the backup ran.
What we're still not super clear on is what got messed up on the original backup such that it hung and the files got orphaned. Maybe a momentary network issue, or something else? Who knows.
So a new backup is running with just one VM, and all looks well. I'll be adding the other ones back, re-adding GRT, etc...just to make sure it works with all features enabled as well. If not, that may point to the issue with the original job. Also, the BE server itself is on this host and there's apparently a possibility the API may have gone into some kind of loop and caused the original issue as well. So that will be another test.
Also, apparently the "resource not snappable" warnings I was seeing in the job logs were a red herring and nothing to be concerned about.
Thanks again to VJware, and I'll be checking back in here after the tests for other people's future reference. :)
