Re: In-placre persistance change of a relation

On Fri, Dec 17, 2021 at 02:33:25PM +0000, Jakub Wartak wrote:
> >  Justin wrote:
> > On Fri, Dec 17, 2021 at 09:10:30AM +0000, Jakub Wartak wrote:
> > > As the thread didn't get a lot of traction, I've registered it into current
> > commitfest
> > https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcommitf
> > est.postgresql.org%2F36%2F3461%2F&data=04%7C01%7CJakub.Wartak%
> > 40tomtom.com%7Cb815e75090d44e20fd0a08d9c15b45cc%7C374f80267b544a
> > 3ab87d328fa26ec10d%7C0%7C0%7C637753420044612362%7CUnknown%7CT
> > WFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXV
> > CI6Mn0%3D%7C3000&sdata=0BTQSVDnVPu4YpECKXXlBJT5q3Gfgv099SaC
> > NuBwiW4%3D&reserved=0 with You as the author and in 'Ready for
> > review' state.
> > 
> > The patch is failing:
> [..] 
> > I think you ran "make check", but should run something like this:
> > make check-world -j8 >check-world.log 2>&1 && echo Success
> 
> Hi Justin,
> 
> I've repeated the check-world and it says to me all is ok locally (also with --enable-cassert --enable-debug , at
leaston Amazon Linux 2) and also installcheck on default params seems to be ok  
 
> I don't seem to understand why testfarm reports errors for tests like "path, polygon, rowsecurity" e.g. on
Linux/graviton2and FreeBSD while not on MacOS(?) . 
 
> Could someone point to me where to start looking/fixing?

Since it says this, it looks a lot like a memory error like a use-after-free
 - like in fsync_parent_path():

 CREATE TABLE PATH_TBL (f1 path);
+ERROR:  could not open file <....>  Pacific": No such file or directory

I see at least this one is still failing, though:
time make -C src/test/recovery check