Hi Kirill, thanks for looking into this!
> On 20 Aug 2025, at 12:19, Kirill Reshke <reshkekirill@gmail.com> wrote:
>
> + /*
> + * We might have filled this offset previosuly.
> + * Cross-check for correctness.
> + */
> + Assert((*offptr == 0) || (*offptr == offset));
>
> Should we exit here with errcode(ERRCODE_DATA_CORRUPTED) if *offptr !=
> 0 and *offptr != offset?
No, we should not exit. We encountered inconsistencies that we are fully prepared to fix. But you are right - we should
betteremit WARNING with XX001.
> + /* Read and adjust next page */
> + next_slotno = SimpleLruReadPage(MultiXactOffsetCtl, next_pageno, true, next);
> + next_offptr = (MultiXactOffset *)
> MultiXactOffsetCtl->shared->page_buffer[next_slotno];
> + next_offptr[next_entryno] = offset + nmembers;
>
> should we check the value of next_offptr[next_entryno] to be equal to
> zero or offset + nmembers ? Assert or
> errcode(ERRCODE_DATA_CORRUPTED) also.
Yes, we'd better WARN user here.
Thanks for your valuable suggestions. I'm not sending new version of the patch, because I'm waiting input on overall
designfrom Alvaro or any committer willing to fix this. We need to figure out if this radical approach is acceptable to
backpatch.I do not see other options, but someone might have more clever ideas.
Best regards, Andrey Borodin.
