Andrey Borodin
Andrey Borodin
Hi! Seems like you have extracted to ```/data```, but starting ```pg_ctl -D /data/main start```
But I see the point how to make WAL-G better here. We can detect PGDATA from environment end print warning that it is not the place where backup extraction taking...
yes, lack of pg_contol file means that backup was not resotred properly we intentionally extract pg_control file last to prevent attempt to start incompletely restored DB
@456vv it's not a bug, it's a feature. Incomplete restores results in absence of pg_control file.
@456vv I do not understand. What's the problem you are trying to solve?
Did you have any messages in ```wal-g backup-fetch``` log?
hmm..this seems like a windows-related bug? Can you send a PR to fix this?
The biggest challenge of this issue is defining clear and unambiguous interface. Our ```delete``` command is already overloaded by arguments. Maybe we could make something like "retention config" and apply...
Hi! By hardcoded default WAL-G tries [15 times ](https://github.com/wal-g/storages/blob/687adc75037532b87f5446cdc0aee50402b107cc/s3/folder.go#L38). Probably, this setting worth to be configured. But if storage answers 500 too many times - we have no chance to...
BTW PR with native Backblaze support would be welcomed. It's not that difficult, just a bunch of methods like PutObject\GetObject\DeleteObject...