On Thu, Oct 30, 2014 at 03:52:01PM -0400, Tom Lane wrote:
> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> > Robert Haas wrote:
> >> A colleague at EnterpriseDB today ran into a situation on PostgreSQL
> >> 9.3.5 where the server went into an infinite loop while attempting a
> >> VACUUM FREEZE; it couldn't escape _bt_getstackbuf(), and it couldn't
> >> be killed with ^C. I think we should add a check for interrupts into
> >> that loop somewhere;
>
> > Our design principle in this area is that all loops should have
> > CHECK_FOR_INTERRUPTS() calls somewhere, so that even if data is horribly
> > corrupted you can get out of it.
>
> FWIW, I concur with Alvaro that adding a CHECK_FOR_INTERRUPTS() needn't
> require much discussion.
+1
> Given the lack of prior complaints about this
> loop, I'm not sure I see the need to work harder than that; corruption
> of this sort must be quite rare.
Looks like _bt_getstackbuf() is always called with some buffer lock held, so
CHECK_FOR_INTERRUPTS() alone would not help:
http://www.postgresql.org/message-id/flat/16519.1401395152@sss.pgh.pa.us