Dean Henze

> Some experimetation with the workflow I described above reveals that this is relatively efficient for 100 to 1000 granules (6 s to 3'ish minutes) but runs into memory issues...

Being my first Github issue, not sure how this works. Should I close the issue for now? Sounds like you already have plans on getting to it.

> I think we should keep it open until one of the proposed solutions works for this use case. I tried out your solution (auth_env() function) for both the toy...

I am running into errors when scaling, but I think those are connected with worker memory release issues, rather than the one in this ticket.

> FWIW, I managed to run through the whole MUR collection in under 10 minutes using a small 14GB RAM instance! the trick was to use the implicit login() and...

Another update, garbage collection does not seem to be cutting it for my more complicated case. I have the Dask task graph up and watch the memory per work slowly...

> If you share your notebook maybe we can take a look at what's happening. Here is the more complicated case (I think that's what you are referring to?) [https://github.com/DeanHenze/podaac-dpub/blob/main/cowvr_animation.ipynb](https://github.com/DeanHenze/podaac-dpub/blob/main/cowvr_animation.ipynb)....

I ran another test with the toy example to see if garbage collection and memory cleanup were working, and I am not sure they are. I think it's just that...

@betolink , totally understood and agreed, this seems far out of earthaccess's field of view, and not expecting any solutions from you! Just figured I'd follow up with you since...

Hi @owenlittlejohns when you say "removing the items" are you referring to cookbook notebooks? I think there should only be one or so, in which case you can remove them....