David A. Wheeler
We’ve just learned that someone had created a project badge entry on a *test* tier (master), with the intent of using that as the main information source. Since that’s not what the test tiers are for, that did not work out well. We occasionally copy data from the production site to the test tiers (master and staging), and we only learned about this situation after the data had already been overwritten (eek!). We have many backups to prevent the loss of *production* data, but no one had considered the *test* tier data as something that needed backups, so some data was lost from that project’s badge entry. The data can be recreated, but that’s not something I want to do often (!).
So: If you know of a badge entry that’s starting on master.bestpractices.coreinfrastructure.org or staging.bestpractices.coreinfrastructure.org, please save the work (say as a JSON file, just add “.json” to the project URL), and let us help you move that data. Please do NOT put “real” data on a test tier!!
We’re also taking these steps to prevent a recurrence:
1. We’ve turned on a warning message on non-production tiers (including master and staging) for when you try to create a new project, view a project, or edit a project (thanks to Jason Dossett, who implemented this). The warning is at the top and will clearly state that this is a test tier. That should make a recurrence unlikely.
2. We’ve turned on daily backups for master and staging. Normally that’s pointless, but if someone does this again & lets us know within a few days, we should be able to recover the data. Thus, even if this problem recurs, the backups should reduce the probability or impact of data loss.
--- David A. Wheeler