Re: Proposal for changes to recovery.conf API

From: David Steele <david(at)pgmasters(dot)net>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Proposal for changes to recovery.conf API
Date: 2016-09-06 11:52:45
Message-ID: 4bbfc0df-ac7a-26dd-96fe-06b79bd6d0e1@pgmasters.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 9/6/16 4:56 AM, Petr Jelinek wrote:
> On 06/09/16 07:18, Abhijit Menon-Sen wrote:
>>
>> Do we want something like this (easy to implement and document, perhaps
>> not especially convenient to use):
>>
>> recovery_target = 'xid' # or 'time'/'name'/'lsn'/'immediate'
>> recovery_target_xid = xxx? # the only setting we care about now
>> recovery_target_otherthings = parsed_but_ignored
>>
>> Or something like this (a bit harder to implement):
>>
>> recovery_target = 'xid:xxx' # or 'time:xxx' etc.
>>
>
> Personally, I never liked the fact that we have several config variables
> for this and then the last one is chosen (even when it was in
> recovery.conf). We support one recovery_target at a time so it would
> make sense to have single config option for it IMHO.
>
> So +1 on the recovery_target = 'xid:xxx' idea.

I would rather not combine the type and target into a single field - how
about having two fields:

recovery_target_type = 'xid|time|name|immediate'
recovery_target = 'value'

recovery_target would be ignored when recovery_target_type = 'immediate'.

--
-David
david(at)pgmasters(dot)net

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2016-09-06 11:59:06 Re: condition variables
Previous Message Tom Lane 2016-09-06 11:34:23 Re: Postgres abort found in 9.3.11