With server configuration and application functional we cannot conclude postgresql.conf, application behavior need to consider(mandatory)
but, can give some ideal way of calculation
1) shared_buffers can be 7-8% of database size under condition having proper indexes with size not greater then 10-15% of respective table size,
2) work_mem can be 8mb if database size under 100-120gb with point#1
3) maintanance_work_mem can be 512mb if database size under 100-120gb and any table max size under 0-2GB
4) checkpoint_segments = 16
5) all scan parameters "on"
6) autovacuum = on ##### if transactions on tables are under 0-30% per day
autovacuum_vacuum_scale_factor = 0.2
autovacuum_analyze_scale_factor = 0.1
7) effective_cache_size = 64-96GB
Note: 1) may need changes based on application behavior, specially if sorting and full-table-scans and transactions are very high.
2) regular maintenance is required to control database size and performance
3) only some parameters defined
On Wed, Sep 2, 2015 at 5:46 PM, FattahRozzaq <ssoorruu@xxxxxxxxx> wrote:
Dear all,
I have server as below:
- OS Debian 7.0 (Wheezy)
- CPU 24 cores
- RAM 128GB
- SSD 128GB
- 1 Gigabit Ethernet
I need to tune the Postgres with hardware specification above for my
Messaging/Messenger/Chatting Application that will run separately in
another server.
Can you give me the recommended configuration (postgresql.conf ?) for
my requirement above?
Big thanks in advance,
FattahRozzaq
--
Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
