Re: [Todo item] Add entry creation timestamp column topg_stat_replication

Поиск
Список
Период
Сортировка
От Michael Paquier
Тема Re: [Todo item] Add entry creation timestamp column topg_stat_replication
Дата
Msg-id 20181204035625.GQ3423@paquier.xyz
обсуждение исходный текст
Ответ на Re: [Todo item] Add entry creation timestamp column topg_stat_replication  (Michael Paquier <michael@paquier.xyz>)
Ответы Re: [Todo item] Add entry creation timestamp column topg_stat_replication  (Michael Paquier <michael@paquier.xyz>)
Список pgsql-hackers
On Fri, Nov 30, 2018 at 05:54:15PM +0900, Michael Paquier wrote:
> Looks pretty to me at quick glance, unfortunately I have not spent much
> time on it, particularly testing it.
>
> +    <row>
> +     <entry><structfield>reply_time</structfield></entry>
> +     <entry><type>timestamp with time zone</type></entry>
> +     <entry>Send time of last message received from WAL
> receiver</entry>
> +    </row>
> This is a bit confusing as this counts for the last *standby* message
> ('r'), and not the last feedback message ('h').
>
> +   /*
> +    * timestamp of the latest message, reported by standby server
> +   */
> +   TimestampTz replyTime;
>
> A small indentation problem, not a big deal.

I have been looking at this patch more in-depth, and you missed one
critical thing: hot standby feedback messages also include the timestamp
the client used when sending the message, so if we want to track the
latest time when a message has been sent we should track it as much as
the timestamp from status update messages.

Fixing that and updating a couple of comments and variables, I am
finishing with the attached.  Thoughts?

(The catversion bump is a self-reminder, don't worry about it.)
--
Michael

Вложения

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Peter Geoghegan
Дата:
Сообщение: Re: Index Skip Scan
Следующее
От: "Ideriha, Takeshi"
Дата:
Сообщение: RE: idle-in-transaction timeout error does not give a hint