On Mon, May 9, 2011 at 3:32 PM, Leonardo Francalanci <m_lists@yahoo.it> wrote:
>> > I have an index on a timestamp value that is inserted, for 90%
>> > of the inserts, in increasing order. No updates, no deletes on the
>> > table (appends only).
>>
>> The bit about "increasing order" is a red herring here. If you have
>> no updates, then you can leave the FILLFACTOR alone.
>>
>> FILLFACTOR controls how much extra room there is in the way the table
>> is stored, so that if a row is UPDATEd it might be possible to store
>> the row in the same disk page. This alleviates certain pathological
>> conditions with high-UPDATE tables and the way Postgres stores the
>> data (the non-overwriting storage manager).
>
>
> (please add the list when replying to emails)
>
> I'm talking about the index fillfactor, not the table fillfactor...
It will be really useful to see some test results where you alter the
fillfactor and report various measurables.
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services