On Fri, Aug 8, 2014 at 11:43 PM, Fujii Masao <masao.fujii@gmail.com> wrote:
> On Mon, Apr 14, 2014 at 11:40 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Fujii Masao <masao.fujii@gmail.com> writes:
>>> On Tue, Apr 1, 2014 at 1:41 AM, Robert Haas <robertmhaas@gmail.com> wrote:
>>>> Should we try to install some hack around fastupdate for 9.4? I fear
>>>> the divergence between reasonable values of work_mem and reasonable
>>>> sizes for that list is only going to continue to get bigger. I'm sure
>>>> there's somebody out there who has work_mem = 16GB, and stuff like
>>>> 263865a48973767ce8ed7b7788059a38a24a9f37 is only going to increase the
>>>> appeal of large values.
>>
>>> Controlling the threshold of the size of pending list only by GUC doesn't
>>> seem reasonable. Users may want to increase the threshold only for the
>>> GIN index which can be updated heavily, and decrease it otherwise. So
>>> I think that it's better to add new storage parameter for GIN index to control
>>> the threshold, or both storage parameter and GUC.
>>
>> Yeah, -1 for a GUC. A GIN-index-specific storage parameter seems more
>> appropriate.
>
> The attached patch introduces the GIN index storage parameter
> "PENDING_LIST_CLEANUP_SIZE" which specifies the maximum size of
> GIN pending list. If it's not set, work_mem is used as that maximum size,
> instead. So this patch doesn't break the existing application which
> currently uses work_mem as the threshold of cleanup operation of
> the pending list. It has only not to set PENDING_LIST_CLEANUP_SIZE.
>
> This is an index storage parameter, and allows us to specify each
> threshold per GIN index. So, for example, we can increase the threshold
> only for the GIN index which can be updated heavily, and decrease it otherwise.
>
> This patch uses another patch that I proposed (*1) as an infrastructure.
> Please apply that infrastructure patch first if you apply this patch.
>
> (*1)
> http://www.postgresql.org/message-id/CAHGQGwEanQ_e8WLHL25=bm_8Z5zkyZw0K0yiR+kdMV2HgnE9FQ@mail.gmail.com
>
> Regards,
>
> --
> Fujii Masao
Sorry, I forgot to attached the patch.... This time, attached.
Regards,
--
Fujii Masao