On Mon, 28 Oct 2019 16:47:02 +0900 (JST)
Kyotaro Horiguchi <horikyota.ntt@gmail.com> wrote:
> At Fri, 25 Oct 2019 12:28:38 -0400, Tom Lane <tgl@sss.pgh.pa.us> wrote in
> > Jehan-Guillaume de Rorthais <jgdr@dalibo.com> writes:
> > > When investigating for the bug reported in thread "logical replication -
> > > negative bitmapset member not allowed", I found a way to seg fault
> > > postgresql only when cassert is enabled.
> > > ...
> > > I hadn't time to digg further yet. However, I don't understand why this
> > > crash is triggered when cassert is enabled.
> >
> > Most likely, it's not so much assertions that provoke the crash as
> > CLOBBER_FREED_MEMORY, ie the actual problem here is use of already-freed
> > memory.
>
> Agreed.
>
> By the way I didn't get a crash by Jehan's script with the
> --enable-cassert build of the master HEAD of a few days ago.
I am now working with HEAD and I can confirm I am able to make it crash 99% of
the time using my script.
It feels like a race condition between cache invalidation and record
processing from worker.c. Make sure you have enough write activity during the
test.
> FWIW I sometimes got SEGVish crashes or mysterious misbehavor when
> some structs were changed and I didn't do "make clean". Rarely I
> needed "make distclean". (Yeah, I didn't ususally turn on
> --enable-depend..)
I'm paranoid, I always do:
* make distclean
* git reset; git clean -df
* ./configure && make install
Regards,
