So, we've been having some trouble as of late with the resultsdb database being reset and losing all the data out of it.
Turns out that the problem could be traced back to [[http://infrastructure.fedoraproject.org/cgit/ansible.git/commit/?id=ce9a4403241e27e189e62198fbd9be3e7aad9bee | an ansible command in the resultsdb-backend role that was supposed to init the database only once]] (that is only supposed to run if /etc/resultsdb/db-is-init does not exist and touch the file if init_db is run).
I suspect that the root of the particular problem is an ansible bug or syntax change that we didn't detect. Either way, I don't think that we want runs of resultsdb init_db to just drop all the database tables and re-initialize them without warning.
resultsdb init_db
Make it more difficult to clear the database out with really meaning it. Setting priority to high because we can't afford to have these kind of bugs hit us once we go to production.
This ticket is a duplicate of https://pagure.io/taskotron/resultsdb/issue/24
This ticket had assigned some Differential requests: D231
Solved by D231
Metadata Update from @tflink: - Issue tagged with: infrastructure
Login to comment on this ticket.