Re: Possible race condition in pg_mkdir_p()? - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: Possible race condition in pg_mkdir_p()?
Date
Msg-id 20190731044054.GL1577@paquier.xyz
Whole thread Raw
In response to Re: Possible race condition in pg_mkdir_p()?  (Ning Yu <nyu@pivotal.io>)
Responses Re: Possible race condition in pg_mkdir_p()?  (Ning Yu <nyu@pivotal.io>)
List pgsql-hackers
On Wed, Jul 31, 2019 at 12:26:30PM +0800, Ning Yu wrote:
> Could I double confirm with you that you made a clean rebuild after
> applying the patches?  pg_mkdir_p() is compiled as part of libpgport.a,
> and the postgres makefile will not relink the initdb binary
> automatically, for myself I must 'make clean' and 'make' to ensure
> initdb gets relinked.

For any patch I test, I just do a "git clean -d -x -f" before building
as I switch a lot across stable branches as well.  It looks that you
are right on this one though, I have just rebuilt from scratch and I
don't see the failures anymore.
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Ning Yu
Date:
Subject: Re: Possible race condition in pg_mkdir_p()?
Next
From: Amit Kapila
Date:
Subject: Re: POC: Cleaning up orphaned files using undo logs