Re: recovering from "found xmin ... from before relfrozenxid ..."

Hi,

On 2020-07-13 17:12:18 -0400, Robert Haas wrote:
> 1. There's nothing to identify the tuple that has the problem, and no
> way to know how many more of them there might be. Back-patching
> b61d161c146328ae6ba9ed937862d66e5c8b035a would help with the first
> part of this.

Not fully, I'm afraid. Afaict it doesn't currently tell you the item
pointer offset, just the block numer, right? We probably should extend
it to also include the offset...

> 2. In some other, similar situations, e.g. where the tuple data is
> garbled, it's often possible to get out from under the problem by
> deleting the tuple at issue. But I think that doesn't necessarily fix
> anything in this case.

Huh, why not? That worked in the cases I saw.

> Therefore, one of my colleagues has - at my request - created a couple
> of functions called heap_force_kill() and heap_force_freeze() which
> take an array of TIDs. The former truncates them all to dead line
> pointers. The latter resets the infomask and xmin to make the xmin
> frozen. (It should probably handle the xmax too; not sure that the
> current version does that, but it's easily fixed if not.)

xmax is among the problematic cases IIRC, so yes, it'd be good to fix
that.

Greetings,

Andres Freund