First ran in QA, then prod run below.

Note that no SQL snapshot was taken just before this upgrade.


## QA

Date: 2021-11-22 (Pacific)

Testing database migration to v0.5.0 in QA.

Commands and timing:

    git checkout bnewbold-content-scope

    git log | head -n1
    # commit dd00cec4164c1a1c31c8d9cffb92deb2e30b2211

    # as fatcat, in rust directory
    time diesel migration run

    # Running migration 2021-11-17-222046_content_scope
    # real    1m45.095s
    # user    0m0.023s
    # sys     0m0.008s

Hrm, seems like the `ALTER TABLE` resulted in full table scans, which is
unexpected but not too bad.

    time cargo build --release

    # Finished release [optimized] target(s) in 4m 44s
    # real    4m45.105s
    # user    35m35.911s
    # sys     0m57.502s

    time pipenv install --dev --deploy

    # real    0m31.618s
    # user    0m26.830s
    # sys     0m3.873s

Did a test edit to add `content_scope` on a file entity, using web TOML form,
and it worked. Ready for prod!


## Prod

Prep steps:
- merge branch to master
- tag (signed)

Primary Deployment Steps:
- halt all workers on prod1
- git pull
- run diesel SQL migration: time diesel migration run
- cargo build: time cargo build --release
- restart fatcat-api
- pipenv update: time pipenv install --dev --deploy
- restart fatcat-web
- restart all workers on prod1
- deploy to misc-vm, fatcat-prod2-vm, scholar2-vm

Commands:

    git log | head -n1
    # commit a1f8783c7774be7523f7a0207050e09bf5cdb12b

    time diesel migration run
    # real    1m52.029s
    # user    0m0.041s
    # sys     0m0.020s

    time cargo build --release
    # real    3m19.235s
    # user    8m49.656s
    # sys     0m10.654s

    time pipenv install --dev --deploy
    # real    0m13.378s
    # user    1m44.558s
    # sys     0m16.582s

Done, re-deployed elsewhere, and checked status.