Re-run technical metadata for all media items

[peetucket]

After sul-dlss/technical-metadata-service#572 is merged, we should run all media items in batch back through technical metadata generation so the volume levels are available so we can correctly send items to speechToTextWF. Else the work in #1439 will prevent them from working correctly.

Blocked by sul-dlss/technical-metadata-service#572

• Exported all 65k druids
• Break into 10k chunks
• Run each chunk
• druids_1
• druids_2
• druids_3
• druids_4
• druids_5
• druids_6
• druids_7

[peetucket]

from the technical-metadata-service readme, it sounds like we need a list of all media druids in a file called druids.txt on the server in the base of the rails app, then do something like this:

RAILS_ENV=production bundler exec rake techmd:generate_for_moab_list['true']

1. Produce list of druids by filtering for media items and then running a report: https://argo.stanford.edu/report?f%5Bcontent_type_ssim%5D%5B%5D=media&f%5BobjectType_ssim%5D%5B%5D=item
2. Download report as CSV. Open in excel, remove all columns except druid column. Remove header row. Save as txt file.
3. Suggest doing in 10k batches by splitting into multiple txt files.
4. Place txt files on the server. The one currently being run should be called druids.txt and be in the root of the tech md app. scp ~/Downloads/techmd/druids_1.txt dor-techmd-prod-a.stanford.edu:/opt/app/techmd/dor_techmd/current/druids.txt
5. Start the rake task above, which will queue all the jobs. Suggest doing it in a screen to avoid getting disconnected mid-queuing. screen -S queue
6. Monitor queues at https://dor-techmd-prod-a.stanford.edu/queues/busy
7. Monitor HB at https://app.honeybadger.io/projects/68956/faults?q=-is%3Aresolved+-is%3Aignored&sort=last_seen_desc