Re: bgwriter, regression tests, and default shared_buffers settings

Andrew Dunstan <andrew@dunslane.net>

From: Andrew Dunstan <andrew@dunslane.net>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Robert Haas <robertmhaas@gmail.com>, pgsql-hackers@postgresql.org
Date: 2012-07-22T21:30:46Z
Lists: pgsql-hackers

Attachments

On 07/19/2012 10:32 AM, Andrew Dunstan wrote:
>
> On 07/19/2012 10:12 AM, Tom Lane wrote:
>> Robert Haas <robertmhaas@gmail.com> writes:
>>> On Wed, Jul 18, 2012 at 5:30 PM, Andrew Dunstan 
>>> <andrew@dunslane.net> wrote:
>>>> Or we could provide an initdb flag which would set an upper bound on
>>>> shared_buffers, and have make check (at least) use it.
>>> How about a flag that sets the exact value for shared_buffers, rather
>>> than a maximum?  I think a lot of users would like initdb
>>> --shared-buffers=8GB or whatever.
>> That would be significantly harder to deploy in the buildfarm context.
>> We don't know that all the animals are capable of coping with 16MB
>> (or whatever target we settle on for make check) today.
>>
>>
>
> Yeah - unless we allow some fallback things could get ugly. I do like 
> the idea of allowing a settable ceiling on shared_buffers instead of 
> having it completely hardcoded as now.
>
>
>

Here's a draft patch.

cheers

andrew