I'd suggest not trusting these files. See below.
There should be no duplicates when running through the RESCRIPt pipeline. I suspect something is very wrong. Again, you can look through the provenance for those files via QIIME 2 View. You should be able to see what they did to make the QZA files. However, it should not be possible to make a QZA file with duplicate IDs present. Duplicate IDs are checked for during the QZA file generation process. At this stage an error should be raised.
Solution:
There is a work around to make a 138.2 database with your version of QIIME 2. All you need to do is take a more manual approach as outlined in the original tutorial, by clicking the drop-arrow "The gritty details", just under "Hard mode". Just wget the listed files and run the import commands. (the rescript get-silva-data command simply automates these parts for you). Then you are ready to proceed with the rest of the curation pipeline. Skip and choose whichever steps you think appropriate.
FYI, SILVA just recently started to generate some premade QIIME 2 files. Check those out and see if they work.