0.7.290 memory leak during apero_database.py --update

[larnoldgithub]

We moved from the 288 to the 290 for the quicklook processing, as we couldn't fix the issue wit the obj_db.

When I ran apero_database.py --update to rebuild the db from the old folders (tellu and calib) the RAM usage increased
linearly as shown below. The full rebuilding took 33h, with 200,000 files in tellu/.

I got no error messages at the end of the apero_database.py and the new quicklook290 seems to work as expected. I'm catching up with the nights of current run.

[njcuk9999]

I mean to me if the RAM goes down without killing python its not a "leak" of memory - its just a lot being kept open (for some reason) that is then closed at some point. Though this is just semantics, I tried with v0.7 to fix this and close more things but it is a complex problem.

I aim in v0.8 to address this and hope that having all this stuff "open" for so long goes away (and will thus fix the problem with using multiprocessing - where this seems to climb way higher).

[larnoldgithub]

At least I agree that as far as the RAM is large enough it shouldn't be an issue.

Also on the plot above I don't know why RAM usage increases by 10Gb after midnight on Sunday night.