David Steele

We have discussed this feature internally before but do not have any plans to implement it. It would require a huge amount of work for what we consider to be...

> If I can't rely on exit code, on what should I rely to see if verify is ok ? That's a good question. The theory is that the verify...

> But it is a problem in practice - under heavy load, while backup/restore is in progress (simultaneously for many PG clusters that use the same physical storage as repository),...

> We decided to create MVP-patch with --no-wal-range option in our repo. And then make performance testing and compare with --progress-only. This sounds good to me. Obviously there is a...

> Using name no-wal-range causes error I think what we want here is `wal-range`, which defaults to true. Then the user can specify `no-wal-range` to negate it. I think this...

@hilltracer I have had a look through this again. Thank you for the detailed analysis! I generally agree with @pgstef because even the worst numbers look pretty reasonable to me...

> What about --detailed-output name? So --no-detailed-output will be synonym for --progress-only. I think maybe a boolean is not the way to go here. We may have other options we...

How about `--detail=[progress|full]`. It is also easy to tweak the names before commit if we wish to do so.

Sounds like two votes for `detail-level`. I'm OK with that.

> I would just be a bit concerned about this: > > $ pgbackrest info --stanza=ro9pg --detail-level=progress > stanza: ro9pg > status: ok Yeah, I noticed this, too. Within the...