Hi Alexey,
Couple of notes on how the DB conns are handled in the OpenSIPS
1) each process has its own connections, even if to the same DB
2) if multiple modules need to connect to the same DB, from the same
process, the conn is shared (between all modules using the same DB).
Now, depending on the exact opensips revision you have, you may run a
version with a DB bug in clusterer module (this was fixed in the latest
versions). Try removing the clusterer module and see if you have one
connection per process per DB.
Regards,
Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
https://www.opensips-solutions.com
OpenSIPS Summit 2019
https://www.opensips.org/events/Summit-2019Amsterdam/
On 02/26/2019 09:47 AM, Alexey Kazantsev via Users wrote:
Hi list!
OpenSIPS 2.4.4 (x86_64/linux).
CentOS 7.6.1810.
# we increased this value from default "2"
modparam("db_postgres", "max_db_queries", 20)
modparam("db_postgres", "exec_query_threshold", 60000)
modparam("acc", "db_url", "postgres://user:PaSs@pg-server/db")
modparam("avpops", "db_url", "postgres://user:PaSs@pg-server/db")
modparam("dispatcher", "db_url", "postgres://user:PaSs@pg-server/db")
modparam("clusterer", "db_url", "postgres://user:PaSs@pg-server/db")
My colleague noticed an interesting behavior with PG DB connection.
'netstat -tulpan | grep %pg-server-ip%' showed only 1 ESTABLISHED connection.
When we configured 5432 port explicitly for the 'acc' module (which uses the DB
most actively) ...
modparam("acc", "db_url", "postgres://user:PaSs@pg-server:5432/db")
... we saw several dozens of connections.
And the load on the workers also decreased after that.
Is it a bug?
-----------------------------------------------
BR, Alexey
http://alexeyka.zantsev.com/
_______________________________________________
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
