Hello, Hou zj!
> In my test, if the tuple is updated and new tuple is in the same page,
> heapam_index_fetch_tuple should find the new tuple using HOT chain. So, it's a
> bit unclear to me how the updated tuple is missing. Maybe I missed some other
> conditions for this issue.
Yeah, I think the pgbench-based reproducer may also cause page splits in btree.
But we may add an index to the table to disable HOT.
I have attached a reproducer for this case using a spec and injection points.
I hope it helps, check the attached files.
Best regards,
Mikhail.