[Bug] Successful DANDI upload with multiple jobs spawns hanging processes

[CodyCBakerPhD]

Another odd but reproducible issue on certain systems is that after a successful upload to the archive using the NWB GUIDE App, there are several rogue Python processes spawned in the background that take about ~5 minutes to timeout

In those 5 minutes, any attempt to close and relaunch the application will stall out. After the 5 minutes are up, the app can be used again without issue

The actual underlying issue remains a mystery (possibly swapping away from joblib as per previous discussions with Yarik may magically fix the problem?) but in the meantime there are two paths of action to resolve

(a) just wait 5 minutes; as long as this issue is easy to find when/if any user encounters the error, the problem will eventually fix itself

(b) forcing single job usage (not the current default as of 12/13/2023) also fixed the issue by not spawning the bad processes to begin with; @garrettmflynn suggested we patch this in as the new default just to be safe, which will also help to avoid those previous seg fault errors sporadically observed by users

[CodyCBakerPhD]

Action item here is to just swap DANDI code to not use joblib

[CodyCBakerPhD]

Will keep this in mind, but the current solution isn't too bad. Depending on system and internet, it's possible that a single job can move as quickly as multiple if the bandwidth is the bottleneck