Re: LWLOCK_STATS

On Sat, Jan 7, 2012 at 5:24 AM, Simon Riggs <simon@2ndquadrant.com> wrote:
> On Fri, Jan 6, 2012 at 10:24 PM, Robert Haas <robertmhaas@gmail.com> wrote:
>>  Five-minute pgbench run, scale factor 100,
>> permanent tables, my usual config settings.  Somewhat depressingly,
>> virtually all of the interesting activity still centers around the
>> same three locks
>
> We've seen clear evidence that the performance profile changes over
> time, with certain artifacts becoming more prominent.
>
> Running exactly the same tests repeatedly is useful to derive the
> historical perspectives, but we need a wider spread of tests to be
> certain that the work done is generally applicable.
>
> I'd be interested to see results from a 30 minute run, focusing on
> what happens in minutes 10-30, if you have time.

Yeah, that seems like a good test to run.  I do have time, but Nate
Boley's test machine is currently otherwise occupied, so I can't run
that test just now.  I will run it when/if an opportunity presents
itself...

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company