What parameter file settings are you using? I got low CPU utilization on 11.6
with postgis 2.4.
Get BlueMail for Android
On Feb 25, 2020, 3:31 AM, at 3:31 AM, Stefan Duling
wrote:
>Sorry there is a major typor in my previous mail with the postgres
>version. Each time i wrote 10.4 it has to
Yes, please, if you draw a sample of your tables can you still see the
big time difference in your queries? We cannot easily test against
your production data because (a) we don't have it and (b) it would
take us a long time to just prepare it. I would love to fix this
because you seem to have nail
Hi Stefan,
> I haven’t tested the newly released postgis 3.0.1 since there is no easy
installation so far (as long as I see).
If you can create a minimal reproducible example [
https://en.wikipedia.org/wiki/Minimal_working_example ]
then the core developers can easily replicate, debug and fix thi
Sorry there is a major typor in my previous mail with the postgres version.
Each time i wrote 10.4 it has to be 10.12. Corrected version:
Hello everyone,
I just want to report the outcomes of my latest performance research.
Postgres 12.1 and Postgres 12.2 are both restrictively slow in perfor
Hello everyone,
I just want to report the outcomes of my latest performance research.
Postgres 12.1 and Postgres 12.2 are both restrictively slow in performing our
geometry processing queries - no matter what postgis version (3.0.0, 2.5) I use
and how the raw data came into the db. Setting jit
Hi Imre,
thanks for showing the other discussions. It really seems like it’s a very
similar problem.
I tried:
- jit = off
- max_parallel_workers_per_gather = 0
- jit = off with max_parallel_workers_per_gather = 0
- ANALYZE osm_admin_areas; before the query
Sadly those options have not the effe
Hi Stefan,
> I tried Postgis 2.5 on the new setup with Postgres 12.1. I can’t see any
significant differences.
Please re-test with:
- "ANALYZE osm_admin_areas; " AND " jit=off
max_parallel_workers_per_gather=0" parameters
based on similar problem:
https://github.com/openstreetmap/osm2pgsq
Hi Paul,
thanks for your response!
I tried Postgis 2.5 on the new setup with Postgres 12.1. I can’t see any
significant differences.
I also tried to use only one worker per gatherer. Although there is a little
improvement, the query time still is doubled compared to the old setup with
Postgres
If you can cut this down to a smaller, shorter query that shows the same
characteristics, I would like to see it in a profiler to determine if the hot
spots have moved. The explain seems more or less structurally the same, which
leaves a couple possibilities:
- the st_collect running in parallel
Hi everyone,
at the moment I am trying out a new setup for our geo database server.
Primarily I upgraded from Postgres and Postgis to newer versions. Beside of the
memory both databases run on the same machine setups and contain identical
data. Differences are:
OLD SETUP
16GB RAM
Postgres 10.
10 matches
Mail list logo