INTR-341 Fix search breaking during deployment

[CamLamb]

The search currently breaks because we have atomic rebuild enabled and we do 2 update_index calls, one with --schema-only and one without.

How it breaks:

• The current search index has documents
• The first update_index call with --schema-only builds a new index with just the schema changes and then changes the pointer for the search backend to use this new index (This index has NO documents)
• The second update_index call populates the documents and then changes the pointer over when done

How to test this locally:

• Start your local environment on the main branch

• Run ./manage.py update_index_async

• Navigate to some search results (ideally with people in the response)

• Re run the update_index_async command and keep refreshing the search results page

  • If you are (un)lucky you will catch the scenario where some of the indexes have completed and some haven't, resulting in no/partial search results showing

• To test the fix, simply checkout this branch and rerun the above steps 😄

• JIRA ticket referenced in title

• Title is clear and concise

• Description gives any relevant detail

• Tests are up to date

• Documentation is up to date

[marcelkornblum]

I don't remember why we did this in the first place but it seems very much the right change to make :D