Re: replication slot restart_lsn initialization - Mailing list pgsql-hackers

From Andres Freund
Subject Re: replication slot restart_lsn initialization
Date
Msg-id 20150906113215.GB7137@alap3.anarazel.de
Whole thread Raw
In response to Re: replication slot restart_lsn initialization  (Michael Paquier <michael.paquier@gmail.com>)
Responses Re: replication slot restart_lsn initialization  (Michael Paquier <michael.paquier@gmail.com>)
List pgsql-hackers
On 2015-08-15 21:16:11 +0900, Michael Paquier wrote:
> Well, this has taken less time than I thought:
> =# CREATE_REPLICATION_SLOT toto PHYSICAL;
>  slot_name | consistent_point | snapshot_name | output_plugin
> -----------+------------------+---------------+---------------
>  toto      | 0/0              | null          | null
> (1 row)
> =# CREATE_REPLICATION_SLOT toto2 PHYSICAL RESERVE_WAL;
>  slot_name | consistent_point | snapshot_name | output_plugin
> -----------+------------------+---------------+---------------
>  toto2     | 0/0              | null          | null
> (1 row)

Independently I wonder if it's worth splitting the returned columns for
physical/logical slots.

> +      <varlistentry>
> +       <term><literal>RESERVE_WAL</></term>
> +       <listitem>
> +        <para>
> +         Specify that the <acronym>LSN</> for this physical replication
> +         slot is reserved immediately; otherwise the <acronym>LSN</> is
> +         reserved on first connection from a streaming replication client.
> +        </para>
> +       </listitem>

I replaced LSN with WAL here - it's not LSNs that are reserved.

> +opt_reserve_wal:
> +            K_RESERVE_WAL
> +                { $$ = TRUE }
> +            | /* EMPTY */
> +                { $$ = FALSE }
> +

I felt free to add semicolons here ;)


Committed, thanks for the patch.



pgsql-hackers by date:

Previous
From: dinesh kumar
Date:
Subject: Re: [PATCH] SQL function to report log message
Next
From: Magnus Hagander
Date:
Subject: Re: Add support for RADIUS passwords longer than 16 octets