You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The new code only supports POSTGRES_PASSWORD, breaking anything using the previous existing secret logic. It also has an initial conditional followed by an else and then another conditional; the last two segments are swapped. So if someone were to specify an existing secret using Values.digger.postgres.existingSecretName and also set postgres_enabled to false (it now weirdly defaults to true, which probably should not be the default and should definitely be more loudly communicated in the changelog), the chart would render two valueFrom lines, resulting in invalid yaml.
The text was updated successfully, but these errors were encountered:
Without adding too much commentary because I know I'm in a mood from tracking this all down... Seriously. The changes made there are not patch level. "Major version = 0" doesn't mean it's cool to completely break backwards compatibility in multiple places without adding any release notes at all. That PR broke my setup in at least two completely avoidable places, adding only functionality which was already achievable with the pre-PR chart. And, again, zero documentation of all the new values. :/
Not gonna lie our exposure to helmcharts has been limited before digger so we clearly messed up.
We wanted to improve a few of the defaults but we did a bad job and testing for backwards compatability, I will prioritise a fix for these items to make it work out from this point on and document that the latest release has broke some older defaults
The last commit (2cf4b15#diff-204d7782307a1567576db307ef282e33ef9a56ab6024b064960f777651b15acaL35-R49) / #10 removes the ability to use an existing secret to set multiple environment variables and has broken conditional logic.
The code which was replaced worked with a secret like this, generated by db-operator:
The new code only supports
POSTGRES_PASSWORD
, breaking anything using the previous existing secret logic. It also has an initial conditional followed by an else and then another conditional; the last two segments are swapped. So if someone were to specify an existing secret using Values.digger.postgres.existingSecretName and also set postgres_enabled to false (it now weirdly defaults to true, which probably should not be the default and should definitely be more loudly communicated in the changelog), the chart would render two valueFrom lines, resulting in invalid yaml.The text was updated successfully, but these errors were encountered: