Re: Split xlog.c

Поиск
Список
Период
Сортировка
От Heikki Linnakangas
Тема Re: Split xlog.c
Дата
Msg-id ff82d82b-8905-8f94-740d-da8d52baa6e0@iki.fi
обсуждение исходный текст
Ответ на Re: Split xlog.c  (Kyotaro Horiguchi <horikyota.ntt@gmail.com>)
Список pgsql-hackers
On 17/09/2021 06:10, Kyotaro Horiguchi wrote:
>    recoveryStopReason is always acquired but it is used only after
>    archive recovery.  I'm not sure about reason for the variable to
>    live in that wide context.  Couldn't we remove the variable then
>    call getRecoveryStopReason() directly at the required place?

Robert commented on the same thing, see my reply there.

> 0002:
> 
>    heapam.c, clog.c, twophase.c, dbcommands.c doesn't need  xlogrecvoer.h.

Cleaned that up in v7, thanks!

>> XLogRecCtl
> 
>    "Rec" looks like Record. Couldn't we use "Rcv", "Recov" or just
>    "Recovery" instead?

I never made that association before, but now I cannot unsee it :-). I 
changed it to XLogRecoveryCtl.

>> TimeLineID    PrevTimeLineID;
>> TransactionId oldestActiveXID;
>> bool        promoted = false;
>> EndOfWalRecoveryInfo *endofwal;
>> bool        haveTblspcMap;
> 
>    This is just a matter of taste but the "endofwal" looks somewhat
>    alien in the variables.

Changed to "endOfRecoveryInfo".

> 
> xlog.c:
> +void
> +SwitchIntoArchiveRecovery(XLogRecPtr EndRecPtr)
> 
>    Isn't this a function of xlogrecovery.c?  Or rather isn't
>    minRecoveryPoint-related stuff of xlogrecovery.c?

Updating the control file is xlog.c's responsibility. There are two 
different minRecoveryPoints:

1. xlogrecovery.c has a copy of the minRecoveryPoint from the control 
file, so that it knows when we have reached consistency.

2. xlog.c is responsible for updating the minRecoveryPoint in the 
control file, after consistency has been reached.

SwitchIntoArchiveRecovery() is called on the transition.

- Heikki



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Heikki Linnakangas
Дата:
Сообщение: Re: Split xlog.c
Следующее
От: Andres Freund
Дата:
Сообщение: Re: Reduce function call costs on ELF platforms