[Benchmarking] PostGIS performance issues?
pramsey at opengeo.org
Sat Aug 28 16:48:13 EDT 2010
Someone shut it off perhaps? Were Oracle and PgSQL thrashing memory?
If we both tried to take a big enough shared memory segment nasty
things could happen.
On Sat, Aug 28, 2010 at 1:42 PM, Paul Ramsey <pramsey at opengeo.org> wrote:
> I'm actually not finding a postgis running on the db machine... I
> wonder what's up?
> On Sat, Aug 28, 2010 at 3:07 AM, Michael Smith
> <michael.smith at usace.army.mil> wrote:
>> Did you make any runs against Oracle to see if its the database server vs
>> postgis being disk bound?
>> I'll go ahead and do some runs with Oracle to see what I'm getting.
>> On 8/28/10 6:02 AM, "Andrea Aime" <aaime at opengeo.org> wrote:
>>> to explore a hunch this morning I made a few full runs of the
>>> benchmarks against postgis from GeoServer and MapServer fcgi
>>> getting more or less the same bad results: either of them hardly
>>> goes beyond the 6 requests per second.
>>> Looking at the database machine the test is disk bound with
>>> postgres reading less than 10MB/s for all the test duration
>>> (as reported by vmstat 1)
>>> I'm wondering if there might be anything wrong with the
>>> postgres/postgis configuration itself?
>>> The shapefile tests on the linux server are not disk bound
>>> for mapserver and the performance is completely different
>>> (they are disk bound for GS but I'm working on it ;-) )
>> Benchmarking mailing list
>> Benchmarking at lists.osgeo.org
More information about the Benchmarking