SpikeSortingLoader sl.load_spike_sorting() halts when used in loop

[mschart]

Hi guys,
When trying to load spike sorting results for all BWM sessions, the download often halts without progress (can wait very long, say over night). Is there a way to switch off that multi-threading? I suspect that to be related.

    for pid in pids:
        eid, probe = pid
        sl = SpikeSortingLoader(eid=eid, pname=probe, 
                                one=one_online, atlas=ba)
        spikes, clusters, channels = sl.load_spike_sorting()  

It prints this and that doesn't change for many hours, for example. Restarting the download fixes the issue and it quickly downloads as expected. This does not work when having a loop of 350 pids, as then most need to be manually restarted.

Downloading: /mnt/8cfe1683-d974-40f3-a20b-b217cad4722a/FlatIron/zadorlab/Subjects/CSH_ZAD_011/2020-03-23/001/alf/probe01/pykilosort/spikes.amps.9788b83b-ff75-4dbc-8654-ce4fbce36f77.npy Bytes: 119010352
  0%|                                             | 0/119010352 [00:00<?, ?it/s]Downloading: /mnt/8cfe1683-d974-40f3-a20b-b217cad4722a/FlatIron/zadorlab/Subjects/CSH_ZAD_011/2020-03-23/001/alf/probe01/pykilosort/spikes.depths.a47789db-7344-4ce7-885a-4b9f907e202e.npy Bytes: 119010352
150994944it [11:23, 312864.09it/s]                                              
150994944it [11:42, 301432.85it/s]