Re: Improve description of XLOG_RUNNING_XACTS

Поиск
Список
Период
Сортировка
От Masahiko Sawada
Тема Re: Improve description of XLOG_RUNNING_XACTS
Дата
Msg-id CAD21AoAHxHRMcYmDN6XkReHi5o8eCF10O+Q2kKJgunzMv1T6zg@mail.gmail.com
обсуждение исходный текст
Ответ на Re: Improve description of XLOG_RUNNING_XACTS  (Kyotaro Horiguchi <horikyota.ntt@gmail.com>)
Список pgsql-hackers
On Thu, Jul 28, 2022 at 3:24 PM Kyotaro Horiguchi
<horikyota.ntt@gmail.com> wrote:
>
> At Thu, 28 Jul 2022 09:56:33 +0530, Ashutosh Bapat <ashutosh.bapat@enterprisedb.com> wrote in
> > Thanks Masahiko for the updated patch. It looks good to me.
> >
> > I wonder whether the logic should be, similar
> > to ProcArrayApplyRecoveryInfo()
> >  if (xlrec->subxid_overflow)
> > ...
> > else if (xlrec->subxcnt > 0)
> > ...
> >
> > But you may ignore it.
>
> Either is fine if we asuume the record is sound, but since it is
> debugging output, I think we should always output the information *for
> both* . The following change doesn't change the output for a sound
> record.
>
> ====
>         if (xlrec->subxcnt > 0)
>         {
>                 appendStringInfo(buf, "; %d subxacts:", xlrec->subxcnt);
>                 for (i = 0; i < xlrec->subxcnt; i++)
>                         appendStringInfo(buf, " %u", xlrec->xids[xlrec->xcnt + i]);
>         }
> -       else if (xlrec->subxid_overflow)
> +       if (xlrec->subxid_overflow)
>                 appendStringInfoString(buf, "; subxid overflowed");
> ====

Do you mean that both could be true at the same time? If I read
GetRunningTransactionData() correctly, that doesn't happen.

Regards,

-- 
Masahiko Sawada
EDB:  https://www.enterprisedb.com/



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Fujii Masao
Дата:
Сообщение: Re: Refactoring postgres_fdw/connection.c
Следующее
От: Alvaro Herrera
Дата:
Сообщение: Re: LogwrtResult contended spinlock