Re: v12.0: segfault in reindex CONCURRENTLY

On 2019-Oct-13, Justin Pryzby wrote:

> On Sun, Oct 13, 2019 at 03:10:21PM -0300, Alvaro Herrera wrote:
> > On 2019-Oct-13, Justin Pryzby wrote:
> > 
> > > Looks like it's a race condition and dereferencing *holder=NULL.  The first
> > > crash was probably the same bug, due to report query running during "reindex
> > > CONCURRENTLY", and probably finished at nearly the same time as another locker.
> > 
> > Ooh, right, makes sense.  There's another spot with the same mistake ...
> > this patch should fix it.
> 
> I would maybe chop off the 2nd sentence, since conditionalizing indicates that
> we do actually care.
> 
> +                        * If requested, publish who we're going to wait for.  This is not
> +                        * 100% accurate if they're already gone, but we don't care.

True.  And we can copy the resulting comment to the other spot.

(FWIW I expect the crash is possible not just in reindex but also in
CREATE INDEX CONCURRENTLY.)

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services