Simplify command for PG dump generator in CI, fail early if seeding failed and document troubleshooting
What
- Update CI command for
bin/generate-pg-dump $options
to simplify maintenance - Add general troubleshooting steps when PG Dump creation job fails and output it in Slack
- Raise seeding error early if
docker exec
fails - Posting Slack report only it
report_slack
passed
Why
Example - https://gitlab.com/gitlab-org/quality/pg-dump-generator/-/pipelines/1195464880 - as seen the pipeline passed however seeding failed. This happened because if Docker exec can find the container and run the command, it returns a success status, even if the command it runs inside the container fails. So added workaround to output exit code and track it.
Also related to Monitor multi-version upgrades test stability (gitlab-org/quality/quality-engineering/team-tasks#1922 - closed) for documenting troubleshooting.
Edited by Nailia Iskhakova