Interesting point you made about the read to write ratio of 1 to 15.
How frequently will you be adding new entities or in the case of storing the
customers in one database table, how frequently will you be adding new
objects of a certain entity type. How many entity types do you foresee
existing? i
What types of journaling on each fs?
On Fri, Jun 4, 2010 at 1:26 PM, Jon Schewe wrote:
> On 6/4/10 9:33 AM, Andres Freund wrote:
> > On Friday 04 June 2010 16:25:30 Tom Lane wrote:
> >
> >> Andres Freund writes:
> >>
> >>> On Friday 04 June 2010 14:17:35 Jon Schewe wrote:
> >>>
> XFS (log
UFS2 w/ soft updates on FreeBSD might be an interesting addition to the list
of test cases
On Fri, Jun 4, 2010 at 9:33 AM, Andres Freund wrote:
> On Friday 04 June 2010 16:25:30 Tom Lane wrote:
> > Andres Freund writes:
> > > On Friday 04 June 2010 14:17:35 Jon Schewe wrote:
> > >> XFS (logbufs
Is this a bulk insert? Are you wrapping your statements within a
transaction(s)?
How many columns in the table? What do the table statistics look like?
On Fri, Jun 4, 2010 at 9:21 AM, Michael Gould <
mgo...@intermodalsoftwaresolutions.net> wrote:
> In my opinion it depends on the application,
Agree with Tom on his point about avoidance of cost param adjustments to fit
specific test cases.
A few suggestions...as I assume you own this database...
- check out pg_statio_user_tables - optimize your cache hit ratio on blocks
read...different time durations... pg_stat_bgwriter (read from a scr
