Michael Paquier wrote: > On Sun, Oct 1, 2017 at 10:25 AM, Peter Geoghegan <p...@bowt.ie> wrote: > > On Thu, Sep 28, 2017 at 1:20 PM, Alvaro Herrera <alvhe...@alvh.no-ip.org> > > wrote: > >> Maybe what this means is that we need to do both Dan's initially > >> proposed patch (or something related to it) apart from the fixes already > >> pushed. IOW we need to put back some of the "tupkeep" business ... > > > > I took the time to specifically check if that would fix the problem. > > Unfortunately, it did not. We see exactly the same problem, or at > > least amcheck/REINDEX produces exactly the same error. I checked both > > Dan's original update_freeze.patch, and your revision that retained > > some of the "tupkeep" stuff, > > 0002-Don-t-freeze-recently-dead-HOT-tuples, which you posted on > > September 6th. > > I did not take the time to dig into that more than two hours, but my > first feeling is that some race condition is going on with the heap > pruning.
I'll look into this on Monday. I found out yesterday that the problematic case is when HTSV returns HEAPTUPLE_DEAD and the HOT tests return true. I haven't yet figured if it is one of those specifically, but I suspect it is the IsHotUpdated case. -- Álvaro Herrera https://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
