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 nature of deploying from GitHub checkouts makes it confusing when changes in conf/ unintentionally are rsync'ed to production as part of a deploy. Ran into an issue where we did a deploy to prod from a machine with an old version of master. Took a while to figure out the conf/ was stale, because normally all changes come from GitHub ignoring the local machine. The difference in deploy conf/ versus everything else is confusing.
In this case, it just broke the deploy. In a worse case, it would deploy something broken.
The text was updated successfully, but these errors were encountered:
IMO, this is an user error. It is convenient to deploy small changes to your conf/ without having to commit them to github. Sometimes, it takes a couple of attempts before you get something to work and committing those changes would not be great.
We could add a warning if you have local changes or a non-pulled local copy. That warning could be an error unless you explicitly give a flag to allow it. Especially on production.
The nature of deploying from GitHub checkouts makes it confusing when changes in conf/ unintentionally are rsync'ed to production as part of a deploy. Ran into an issue where we did a deploy to prod from a machine with an old version of master. Took a while to figure out the conf/ was stale, because normally all changes come from GitHub ignoring the local machine. The difference in deploy conf/ versus everything else is confusing.
In this case, it just broke the deploy. In a worse case, it would deploy something broken.
The text was updated successfully, but these errors were encountered: