On Thu, Mar 04, 2021 at 08:03:25PM +0530, Amit Kapila wrote: > I think here we have a slight timing thing which is if the checkpoint > happens before restart then the problem won't occur, not sure but > maybe increase the checkpoint_timeout as well to make it reproducible.
Thanks for providing more details here. So, this also means that you would need a TAP test for that stuff anyway as this relies on a restart, which is what your patches on the other thread are doing, anyway. -- Michael
signature.asc
Description: PGP signature
