scripts/scrub-allura-data.py
does not remove all sensitive data. We should drop these collections:
And check for further collections that should be dropped too. Also consider our internal SF collections and if any need to be scrubbed as well.
The user collection should scrub all fields that aren't explicitly allowed (possible in a mongo query? perhaps copy partial docs to a new collection and rename). Specific examples we're currently missing: password field, and tool_data (for github import tokens).
Afterwards, https://trac.geek.net/trac/siteops/ticket/51121 can be unblocked
Diff: