Re: undetected deadlock in ALTER SUBSCRIPTION ... REFRESH PUBLICATION

On Thu, Dec 7, 2023 at 11:21 AM Shlok Kyal <shlok(dot)kyal(dot)oss(at)gmail(dot)com> wrote:
>
> > I mean to commit the open transaction at the below place in
> > wait_for_relation_state_change()
> >
> > wait_for_relation_state_change()
> > {
> > ...
> > -- commit the xact
> > WaitLatch();
> > ...
> > }
> >
> > Then start after the wait is over. This is just to test whether it
> > improves the difference in regression test timing.
>
> I tried the above approach and observed that the performance of this
> approach is nearly same as the previous approach.
>

Then let's go with the original patch only. BTW, it took almost the
same time (105 wallclock secs) in my environment (CentOs VM) to run
tests in src/test/subscription both with and without the patch. I took
a median of five runs. I have slightly adjusted the comments and
commit message in the attached. If you are fine with this, we can
commit and backpatch this.

--
With Regards,
Amit Kapila.