Closed hhff closed 6 years ago
Hi,
Since you can reproduce this using heroku pg:copy
, I think this is an issue with Heroku CLI (and maybe with arguments it passes to pg_restore
). Our remote environment restore command is a wrapper around Heroku CLI's pg:backups
, so it doesn't do anything beyond get the backup URL and set the destination.
If you open a support ticket with Heroku and the end result is that there's some extra arguments that need to be passed to carry PG functions over, etc. in a restore and we need to include or accommodate those in our restore command, I'd be happy to review a PR that adds those flags.
Until you can successfully move this data over with Heroku CLI using either pg:backups:restore
or pg:copy
, we're not going to be able to fix this issue in Parity.
Thanks, @geoffharcourt !
In case anyone else stumbles across this - I realized this morning that my staging database was on Heroku's hobby-dev
plan, and my production backup was blowing out it's row limit.
Not sure if the issue was with the row limit (heroku claims not to block inserts until 7 days after the violation) or with hobby-dev not importing sequencing functions, but either way, upgrading staging to hobby-basic
fixed this issue 👍
Thank you so much for the hard work!
What command did you execute?
FYI - on staging:
and on production:
It appears that
@default_function
is not being persisted with the database backup.I also observe the same issue when using
pg:copy
:heroku pg:copy production_app::DATABASE_URL DATABASE_URL -a staging_app
What did you expect to happen?
A post would be created.
What actually happened?
This error:
Some information about your installation
What's your operating system? MacOS
What's the output of
which development
,which staging
,which production
? Parity has had multiple installation channels, and it's not uncommon for an old version to be somewhere else in your path.brew list parity
output?