Re: HOT vs freezing issue causing "cannot freeze committed xmax"

2020-07-24 Thread Andres Freund
Hi, On 2020-07-24 11:06:58 -0400, Robert Haas wrote: > On Thu, Jul 23, 2020 at 2:10 PM Andres Freund wrote: > > In the case the HOT logic triggers, we'll call > > heap_prepare_freeze_tuple() even when the tuple is dead. > > I think this is very bad. I've always been confused about these > comment

Re: HOT vs freezing issue causing "cannot freeze committed xmax"

2020-07-24 Thread Robert Haas
On Thu, Jul 23, 2020 at 2:10 PM Andres Freund wrote: > In the case the HOT logic triggers, we'll call > heap_prepare_freeze_tuple() even when the tuple is dead. I think this is very bad. I've always been confused about these comments, but I couldn't quite put my finger on the problem. Now I think

HOT vs freezing issue causing "cannot freeze committed xmax"

2020-07-23 Thread Andres Freund
Hi, In a development branch of mine Thomas / the CF bot found a relatively rare regression failures. That turned out to be because there was an edge case in which heap_page_prune() was a bit more pessimistic than lazy_scan_heap(). But I wonder if this isn't an issue more broadly: The issue I am c