"Tom Lane" <[EMAIL PROTECTED]> writes:
> "Heikki Linnakangas" <[EMAIL PROTECTED]> writes:
>> * mmap the WAL segments, instead of using the slru buffers.
>
> This one's almost certainly a nonstarter, for lack of any portable way
> to control when writes happen.
I think mlock and msync(MS_SYNC) oug
"Heikki Linnakangas" <[EMAIL PROTECTED]> writes:
> We mustn't write data pages before the WAL hits the disk, but we don't
> have any such limitation for WAL.
Um, you're right, I was remembering discussions about trying to mmap
data files. Still, googling "msync performance" leaves me full of
conc
Tom Lane wrote:
> "Heikki Linnakangas" <[EMAIL PROTECTED]> writes:
>> * mmap the WAL segments, instead of using the slru buffers.
>
> This one's almost certainly a nonstarter, for lack of any portable way
> to control when writes happen.
For flushing, there's msync, which I believe is quite porta
"Heikki Linnakangas" <[EMAIL PROTECTED]> writes:
> * mmap the WAL segments, instead of using the slru buffers.
This one's almost certainly a nonstarter, for lack of any portable way
to control when writes happen.
regards, tom lane
---(end of broadc
Gokulakannan Somasundaram wrote:
>Finally i found some time to look more into the CRC code. The
> time is mostly taken when there is a back-up block in the XLOG structure.
> when it calculates the CRC for the entire block(there is some optimization
> already for the holes), the time is
Hi Heikki,
Finally i found some time to look more into the CRC code. The
time is mostly taken when there is a back-up block in the XLOG structure.
when it calculates the CRC for the entire block(there is some optimization
already for the holes), the time is spent on the CRC macro. I trie
Gokulakannan Somasundaram wrote:
> Hi Tom/ Heikki,
>Thanks for the suggestion. After profiling i got similar results.
> So i am thinking of a design like this to get the performance improvement.
>
> a) We can get one page for insert(during update) and we will hold the write
> lock on
Hi Tom/ Heikki,
Thanks for the suggestion. After profiling i got similar results.
So i am thinking of a design like this to get the performance improvement.
a) We can get one page for insert(during update) and we will hold the write
lock on it, till the page gets filled. In this way,
R
Gokulakannan Somsundaram wrote:
> Again if this full table updates are thought with the OLTP applications in
> mind, then this is not at all a suitable option. This will only benefit the
> people with Data Warehouses.
>
> Expecting some more replies
Start with profiling.
I just ran a quick o
Hi Tom,
Thanks for the feedback. Let me clarify my intention again.
This was thought for improving the performance of the Data Warehousing
applications
Full table updates similar to
"Update dd set n2=n2+1"
When you talked about firing triggers, i looked into the implementation of
triggers a
"Gokulakannan Somsundaram" <[EMAIL PROTECTED]> writes:
> I propose to change this row-by-row approach, when it is a full table
> update. I plan to send a extra flag(which will be set for Full table
> Deletes/Updates). this would make the access method directly acquire the
> exclusive lock and updat
The obvious advantages are
a) Avoidance of one read lock per page
b) One Big write lock instead of multiple write locks.
But as you said, i will do some initial profiling and get back.
Thanks,
Gokul.
On 9/20/07, Heikki Linnakangas <[EMAIL PROTECTED]> wrote:
>
> Gokulakannan Somsundaram wrote:
>
Gokulakannan Somsundaram wrote:
> Hi,
>The Current architecture of Updates in PostgreSQL is
> 1) Make a select query out of update. It involves a READ lock/BUFFER_SHARE
> 2) Get the tupleid
> 3) Goto the buffer containing the tupleid, make a BUFFER_EXCLUSIVE lock on
> it
> 4) update it
> 5) Rep
13 matches
Mail list logo