On Mon, Jul 07, 2025 at 08:00:01AM +0300, Alexander Lakhin wrote:
> I'm sorry for not being accurate enough -- I forgot to mention that I
> replicated the config from culicidae, and now I see that "fsync=on"
> is needed to reproduce the failure for me (though maybe).
Even with that, the situation was stable here with two hours running
the test in a loop.
> I got failures on iterations 3, 5, 1. With your patch applied, I got 100
> iterations passed.
This is good enough for me, so applied. Thanks for double-checking.
--
Michael