On 03.02.2013 08:24, Pavan Deolasee wrote:
On Sun, Feb 3, 2013 at 2:31 AM, Jeff Janes wrote:
I've attached a patch with these changes made. Does this look OK?
Looks good to me. I also repeated pgbench and make check and they work
as expected. I'll add it to the CF and also mark the patch "re
On Sun, Feb 3, 2013 at 2:31 AM, Jeff Janes wrote:
> Hi Pavan,
>
> I get this warning:
> vacuumlazy.c:890: warning: passing argument 6 of 'lazy_vacuum_page'
> makes pointer from integer without a cast
>
> and make check then fails.
>
> I've added '&' to that line, and it now passes make check with
On Sat, Jan 26, 2013 at 11:25 PM, Pavan Deolasee
wrote:
> On Thu, Jan 24, 2013 at 9:31 PM, Jeff Janes wrote:
>> On Thu, Jan 24, 2013 at 1:28 AM, Pavan Deolasee
>> wrote:
>
>>>
>>> Good idea. Even though the cost of pinning/unpinning may not be high
>>> with respect to the vacuum cost itself, but
On Thu, Jan 24, 2013 at 9:31 PM, Jeff Janes wrote:
> On Thu, Jan 24, 2013 at 1:28 AM, Pavan Deolasee
> wrote:
>>
>> Good idea. Even though the cost of pinning/unpinning may not be high
>> with respect to the vacuum cost itself, but it seems to be a good idea
>> because we already do that at othe
On Thu, Jan 24, 2013 at 1:28 AM, Pavan Deolasee
wrote:
> Hi Jeff,
>
> On Thu, Jan 24, 2013 at 2:41 PM, Jeff Janes wrote:
>>
>> lazy_vacuum_page now pins and unpins the vmbuffer for each page it marks
>> all-visible, which seems like a lot of needless traffic since the next
>> vmbuffer is likely t
Hi Jeff,
On Thu, Jan 24, 2013 at 2:41 PM, Jeff Janes wrote:
>
> lazy_vacuum_page now pins and unpins the vmbuffer for each page it marks
> all-visible, which seems like a lot of needless traffic since the next
> vmbuffer is likely to be the same as the previous one.
>
Good idea. Even though the
On Friday, December 7, 2012, Pavan Deolasee wrote:
>
Revised patch attached. This time much less invasive. I added a new
> function heap_page_is_all_visible() to check if the given page is
> all-visible and also return the visibility cutoff xid.
Hi Pavan,
lazy_vacuum_page now pins and unpins t
On Fri, Dec 7, 2012 at 9:21 AM, Pavan Deolasee wrote:
> On Thu, Dec 6, 2012 at 11:31 PM, Tom Lane wrote:
>
>>
>> I think taking a second whack at setting the visibility bit is a fine
>> idea, but let's drop all the rest of this premature optimization.
>>
>
> Fair enough. I thought about doing it
On Friday, December 07, 2012 12:06 AM Robert Haas wrote:
> On Thu, Dec 6, 2012 at 1:01 PM, Tom Lane wrote:
> > I think taking a second whack at setting the visibility bit is a fine
> > idea, but let's drop all the rest of this premature optimization.
>
> +1.
>
> If there's any optimization neede
On Fri, Dec 7, 2012 at 12:05 AM, Robert Haas wrote:
> On Thu, Dec 6, 2012 at 1:01 PM, Tom Lane wrote:
>> I think taking a second whack at setting the visibility bit is a fine
>> idea, but let's drop all the rest of this premature optimization.
>
> +1.
>
> If there's any optimization needed here,
On Thu, Dec 6, 2012 at 11:31 PM, Tom Lane wrote:
>
> I think taking a second whack at setting the visibility bit is a fine
> idea, but let's drop all the rest of this premature optimization.
>
Fair enough. I thought about doing it that way but was worried that an
additional page scan will raise
Robert Haas writes:
> One other thought: I'm wondering if we shouldn't try to push the work
> of setting the all-visible bit into heap_page_prune().
Hm, maybe ...
> But it seems to me that a page can't be all-visible unless there are
> no dead line pointers and no HOT chains of length != 1, and
On Thu, Dec 6, 2012 at 1:01 PM, Tom Lane wrote:
> I think taking a second whack at setting the visibility bit is a fine
> idea, but let's drop all the rest of this premature optimization.
+1.
If there's any optimization needed here, we should try to do it by
remembering relevant details from the
Pavan Deolasee writes:
> So the idea that the patch implements is this. When we scan pages in
> the first phase of vacuum, if we find a page that has all-visible
> tuples but also has one or more dead tuples that we know the second
> phase of vacuum will remove, we mark such page with a special fl
Hi All,
I briefly mentioned this idea in one of the other thread, but starting
a new thread to highlight the point. Today, we set the visibility map
*only* in the first phase of vacuum. This works when the page has no
dead tuples. But the vacuum itself is removing one or more dead tuples
from the
15 matches
Mail list logo
