I tried auto-migrating my v1 postgres app to v2 at about 2pm PDT today July 22, 2023, but the migration failed (timed out waiting for healthy host). Instead of debugging via cli, I canceled.
I am getting alerts saying my db is in read-only mode.
Now the Web UI says the db is “suspended” (grey vertical lines instead of check marks). When viewing the app (url: /apps/{appName}) and I click on “Machines” or “Scale” in the left menu bar, shows “Page not Found”. On my other apps, it takes me to a view with the machines associated.
When I ask to list the machines via cli, it says there are no machines
fly m list -a {appName}
No machines are available on this app
The database is being hosted and my access is able to read/write to it.
Should I:
- create a fresh v2 postgres and perform a data migration?
- retry to automatic migration?
It seems flyio did not disable the read-only mode during the migration rollback. I’m opting for 1/ by manually backing up the existing instance, creating a fresh cluster in v2.