On 2021-Nov-25, Amul Sul wrote:
> In XLogReadRecord(), both the variables being compared have
> inconsistency in the assignment -- one gets assigned from
> state->currRecPtr where other is from RecPtr.
>
> .....
> state->overwrittenRecPtr = state->currRecPtr;
> .....
> state->abortedRecPtr = RecPtr;
> .....
>
> Before the place where assembled flag sets, there is a bunch of code
> that adjusts RecPtr. I think instead of RecPtr, the latter assignment
> should use state->currRecPtr as well.
You're exactly right. I managed to reproduce the problem shown by
buildfarm members, and indeed this fixes it. And it makes sense: the
adjustment you refer to, is precisely to skip the page header when the
LSN is the start of the page, which is exactly the problem we're seeing
in the buildfarm ... except that on lapwing branch REL_11_STABLE, we're
seeing the LSN is off by 0x14 instead of 0x18. That seems very strange.
I think the reason for this is that lapwing has MAXALIGN 4, so
MAXALIGN(sizeof(XLogPageHeaderData)) is 20, not 24 as is the case in the
other failing members.
... checks buildfarm ...
Yeah, all the others in Tom's list are x86-64.
I'm pushing the fix in a minute.
--
Álvaro Herrera PostgreSQL Developer — https://www.EnterpriseDB.com/
"¿Qué importan los años? Lo que realmente importa es comprobar que
a fin de cuentas la mejor edad de la vida es estar vivo" (Mafalda)