AutoQA resultsdb schema

Result

Database schema

Testrun is the representation of the whole test result. The test this testrun represents may implement more phases (one testrun may contain 0..* phases) - All we worry about in test phase is it's name and result).

If one needs to encapsulate more testruns into one (merely imaginary) unit, it's possible to aggregate those testruns using Job (one job may contain 0..* Testruns).

Each test may return additional data (in key-value format) which it needs to store (e.g. compose date, envra,...) - TestrunData is the place to store these.

Default key-values for basic test classes

Installation tests

release_name
compose_date
arch

Repository tests

repo_name
compose_date
arch

Package tests

pkg_name
envr
arch

Recommended:

koji_tag

Specific keyval examples - tests using Bodhi hook:

bodhi_tag/request -- to mark in which state in Bodhi the package was

Specific keyval examples - comparative tests:

counterpart_name
counterpart_envr
counterpart_koji_tag
counterpart_bodhi_tag/request, ...

Next Steps

Package Update Acceptance Test Plan Frontend

Create wiki metadata for testplan & testcases describing it
- Repo sanity
- Upgrade path

Create simple frontend like: https://fedoraproject.org/wiki/File:Package_acceptance_dashboard.png
- where do i get recent builds?
- progress bar:
  - how many tests there are for the testplan? - testplan wiki page
  - what is the progress? - resultsdb - how to get 'the right' results? is 'take the latests tests for envra' way ok? Jobs in autoqa?
- detailed info - detailed state of the testcases (passed/failed/running, detailed phases [if described on the wiki page])

Search