Failed web upgrade

[JammyDodger]

Hello and welcome @Nanovitruvius

There have been a couple of reports of failed upgrades in the past hour or so, and a couple of fixes have now been put in place for those which seemed to have worked for most.

The error "alter schema public owner to discourse;" in your second codeblock cropped up the other day too, and the advice was:

@Ed_S

I’m guessing, but it seems possible that the database was uncleanly shutdown, needs to do a recover, and isn’t quite being given enough time to finish that.

In fact, see this previous thread:

@blattersturm

Whenever this happens, we usually (temporarily, as otherwise Git gets confused when updating the next time) replace the sleep 5 with a few more seconds so that Postgres has more time to settle/recover.

discourse_docker/postgres.template.yml at 4cd6a5d96ab1efee9b6893d781c8af9d3d2f56e2 · discourse/discourse_docker · GitHub

and perhaps see also this much earlier thread:

@Falco

Try increasing the sleep in templates/postgres.template.yml locally. Is your server slow? Database too big?