Simplify db:check-schema CI job
What does this MR do and why?
The db:check-schema CI job times increases over time as more
migrations are introduced to GitLab, and the job applies the migrations
there times in a row for the main, ci, and sec databases. There
are also multiple flavors of this job:
-
db:check-schema-single-db(DECOMPOSED_DB=false) -
db:check-schema-single-db-ci-connection(DECOMPOSED_DB=false,CI_CONNECTION_DB=true) -
db:check-schema-single-db-sec-connection(DECOMPOSED_DB=false,SEC_CONNECTION_DB=true)
This appears to be overkill because the CI job runs the migrations
against an empty database, and all migrations are applied to all
databases in the same way. As a result, we should expect the same result
when running rake db:migrate:main and rake db:migrate:ci, except
which PostgreSQL database was modified.
The job appears to be focused on validating that migrations have properly formatted timestamps and checksums, not on validating the details of the migrations.
We can simplify and save CI minutes by using a single database configuration and removing the other flavors.
Relates to #503221
References
Screenshots or screen recordings
| Before | After |
|---|---|
How to set up and validate locally
MR acceptance checklist
Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.