On Mon, Mar 25, 2019 at 10:29:46AM +0100, Fabien COELHO wrote:
> I have run the non regression tests. I'm not sure of what scenarii are
> covered there, but probably not an interruption in the middle of a fsync,
> specially if fsync is usually disabled to ease the tests:-)
Force the tool to stop at a specific point requires a booby-trap. And
even if fsync is not killed, you could just enforce the tool to stop
once before updating the control file, and attempt a re-run without
the trap, checking if it works at the second attempt, so the problem
is quite independent from the timing of fsync().
--
Michael