On Jun 3, 2011 8:38 PM, "Bruce Momjian" wrote:
>
> I realize we just read the pages from the kernel to maintain sequential
> I/O, but do we actually read the contents of the page if we know it
> doesn't need vacuuming? If so, do we need to?
I dont follow. What's your question?
Tom's final versi
On 03.06.2011 22:16, Bruce Momjian wrote:
I realize we just read the pages from the kernel to maintain sequential
I/O, but do we actually read the contents of the page if we know it
doesn't need vacuuming?
Yes.
If so, do we need to?
Not necessarily, but it allows us to freeze old tuples, a
Heikki Linnakangas wrote:
> On 27.05.2011 16:52, Pavan Deolasee wrote:
> > On closer inspection, I realized that we have
> > deliberately put in this hook to ensure that we use visibility maps
> > only when we see at least SKIP_PAGES_THRESHOLD worth of all-visible
> > sequential pages to take advan
On Fri, May 27, 2011 at 8:40 PM, Greg Stark wrote:
>
> Separately it's a bit strange that we actually have to visit the
> pages. We have all the information we need in the VM to determine
> whether there's a run of 32 vacuum-clean pages. Why can't we look at
> the next 32 pages and if they're all
On Fri, May 27, 2011 at 11:10 AM, Greg Stark wrote:
> It would be nice if the VM had a bit for "all-frozen" but that
> wouldn't help much except in the case of truly cold data. We could
> perhaps keep the frozen data per segment or per VM page (which covers
> a large section of the table) which wo
On Fri, May 27, 2011 at 7:41 PM, Heikki Linnakangas
wrote:
> On 27.05.2011 16:52, Pavan Deolasee wrote:
>>
>> On closer inspection, I realized that we have
>> deliberately put in this hook to ensure that we use visibility maps
>> only when we see at least SKIP_PAGES_THRESHOLD worth of all-visible
On Fri, May 27, 2011 at 7:11 AM, Heikki Linnakangas
wrote:
> Well, as with normal queries, it's usually faster to just seqscan the whole
> table if you need to access more than a few percent of the pages, because
> sequential I/O is so much faster than random I/O.
Well it's not strictly random ac
On Fri, May 27, 2011 at 7:36 PM, Tom Lane wrote:
> Pavan Deolasee writes:
>> My statistical skills are limited, but wouldn't that mean that for a
>> fairly well distributed write activity across a large table, if there
>> are even 3-4% update/deletes, we would most likely hit a
>> not-all-visible
2011/5/27 Cédric Villemain :
> 2011/5/27 Pavan Deolasee :
>> I wonder if we have tested the reasoning behind having
>> SKIP_PAGES_THRESHOLD and the magic number of 32 assigned to it
>> currently. While looking at the code after a long time and doing some
>> tests, I realized that a manual VACUUM wo
On 27.05.2011 16:52, Pavan Deolasee wrote:
On closer inspection, I realized that we have
deliberately put in this hook to ensure that we use visibility maps
only when we see at least SKIP_PAGES_THRESHOLD worth of all-visible
sequential pages to take advantage of possible OS seq scan
optimizations
2011/5/27 Pavan Deolasee :
> I wonder if we have tested the reasoning behind having
> SKIP_PAGES_THRESHOLD and the magic number of 32 assigned to it
> currently. While looking at the code after a long time and doing some
> tests, I realized that a manual VACUUM would always scan first 31
> pages of
Pavan Deolasee writes:
> My statistical skills are limited, but wouldn't that mean that for a
> fairly well distributed write activity across a large table, if there
> are even 3-4% update/deletes, we would most likely hit a
> not-all-visible page for every 32 pages scanned ?
Huh? With a typical
I wonder if we have tested the reasoning behind having
SKIP_PAGES_THRESHOLD and the magic number of 32 assigned to it
currently. While looking at the code after a long time and doing some
tests, I realized that a manual VACUUM would always scan first 31
pages of a relation which has not received an
13 matches
Mail list logo