On Mon, May 6, 2019 at 4:41 PM Thomas Munro <thomas.munro@gmail.com> wrote:
>
> On Mon, May 6, 2019 at 9:26 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> > If so, I am getting it both before and after your patch.
>
> Huh. I thought the only problem here was the phenomenon demonstrated
> by [1]. I'm a bit stumped... if we've closed all the handles in every
> backend before detaching, and then the last to detach unlinks all the
> files first and then the directory, how can we get that error?
>
Yeah, I am also not sure what caused that and I have verified it two
times to ensure that I have not made any mistake. I can try once
again tomorrow after adding some debug messages, but if someone else
can also once confirm the behavior, it would be good.
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com