On Tue, Dec 17, 2013 at 12:04 AM, Jeff Janes <jeff.janes@gmail.com> wrote:
> On Mon, Dec 16, 2013 at 10:02 AM, Robert Haas <robertmhaas@gmail.com> wrote:
>> On Mon, Dec 16, 2013 at 12:02 PM, Jeff Amiel <becauseimjeff@yahoo.com>
>> wrote:
>> Well, the patch was rejected, more or less because people felt it
>> overlapped with pgfincore too much. I don't particularly agree,
>> because pgfincore can't load data into shared buffers and doesn't work
>> on Windows, but other people felt differently.
As far as I can see, it doesn't have the PREWARM_READ mode where
specified pages can be read, also another thing is that as pg_prewarm
uses internal API's, it has certain other advantage as well like for
before PREFETCH, it can ensure whether the block is already in buffer
cache.
>>There was talk of
>> polishing up pgfincore for possible inclusion in contrib, perhaps
>> adding this functionality along the way, but AFAIK there's been no
>> activity on that.
>
>
> It wasn't rejected, it was returned with feedback with generally positive
> reviews. I think the main feedback was that it should provide a
> single-argument overloaded function that takes just the object name and
> applies reasonable defaults for the remaining arguments, for example
> 'main', 'buffer',NULL,NULL. I had thought that the worry about overlap
> with pgfincore was mostly resolved favorably, but perhaps I misread the
> situation.
>
> I'd like to see it revived for 9.4 if you are willing.
I have used pg_prewarm during some of work related to Buffer Management and
other performance related work. It is quite useful utility.
+1 for reviving this patch for 9.4
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com