On 9/1/21, 10:06 AM, "Andres Freund" <andres@anarazel.de> wrote:
> On 2021-09-01 15:01:43 +0900, Fujii Masao wrote:
>> Thanks for clarifying that! Unless I misunderstand that, when recovery ends
>> at a partially-flushed segment-spanning record, it sets
>> XLP_FIRST_IS_ABORTED_PARTIAL in the next segment before starting writing
>> new WAL data there. Therefore XLP_FIRST_IS_CONTRECORD or
>> XLP_FIRST_IS_ABORTED_PARTIAL must be set in the next segment following
>> a partially-flushed segment-spanning record. If none of them is set,
>> the validation code in recovery should report an error.
>
> Right. With the small addition that I think we shouldn't just do this for
> segment spanning records, but for all records spanning pages.
This approach seems promising. I like that it avoids adding extra
work in the hot path for writing WAL. I'm assuming it won't be back-
patchable, though.
Nathan