Bruce Momjian <br...@momjian.us> writes: > Will people now always get a clear error on failure? Crazy idea, but > could we have initdb or postmaster start test this?
It'd require adding quite a lot of cycles: I'd guess you'd need to
create/drop a hundred or a thousand files to be sure of seeing the
problem.
Another problem is that even if the primary data directory is okay,
a tablespace could be created on a filesystem that's not okay.
I doubt that it's worth it.
regards, tom lane
