David Steele

> set audit logging details at group role level and have them inherited by member user roles I think this would be better implemented in PostgreSQL itself, but as I...

I wonder why we can't just add the missing cursor to the condition, rather than making this exception for superuser?

@sfrost Did you look at the attached backtrace? Looks pretty good to me. In any case, I'm pretty sure I know what the issue is -- just not sure about...

Though, it would be better with a line number in `memContextFree()`.

I have been looking into this issue over the last few weeks (since this is the third report) but I have been unable to reproduce. My guess is that there...

We hope that 0f7b6a3344b in 2.41 will solve this issue. It's not perfect because we are avoiding the segfault rather than determining the cause, but at least it should allow...

Believe it or not, this is a bit harder than it sounds. Since there's no overall manifest each file needs to be stat'd to get the actual size and this...

> Why saving to backup.info.copy.pgbackrest.tmp; and later renaming this to backup.info.copy. This is a standard way on Posix to do atomic writes (Postgres does the same). If we wrote to...

The problem is that if you don't write to a tmp file, by the time you know you have truncated data the old file is already gone.

And if the system crashes in middle of that -- what then? What is on disk at that point?