I use multicorn.ldapfdw.LdapFdw.
If I have just connected (psql) to the database, then the first query to the external table takes a long time - 6-7 seconds. After that, any similar request takes 20ms. I disconnect from the server, reconnect. Again the first query to the outer table is 6-7 seconds. How can you speed up the first request so that the extension is already loaded before any new connection? I tried to specify in the configuration file postgresql.conf options:
shared_preload_libraries = 'multicorn'
local_preload_libraries = 'multicorn'
session_preload_libraries = 'multicorn'
and restart postgresql.
Nothing changed.
load 'multicorn' - doesn't help either.
_The foreign data wrapper associated to a table is instantiated on a per-process basis, and it happens when the first query is run against it.
Usually, postgresql server processes are spawned on a per-connection basis.
During the life time of a server process, the instance is cached. That means that if you have to keep references to resources such as connections, you should establish them in the init method and cache them as instance attributes._
I am not familiar with python. What exactly needs to be written in the _init method to have a fast first request?
Hi!
I use multicorn.ldapfdw.LdapFdw. If I have just connected (psql) to the database, then the first query to the external table takes a long time - 6-7 seconds. After that, any similar request takes 20ms. I disconnect from the server, reconnect. Again the first query to the outer table is 6-7 seconds. How can you speed up the first request so that the extension is already loaded before any new connection? I tried to specify in the configuration file postgresql.conf options: shared_preload_libraries = 'multicorn' local_preload_libraries = 'multicorn' session_preload_libraries = 'multicorn' and restart postgresql. Nothing changed. load 'multicorn' - doesn't help either.
https://multicorn.readthedocs.io/en/latest/implementing-tutorial.html
Foreign Data Wrapper lifecycle
_The foreign data wrapper associated to a table is instantiated on a per-process basis, and it happens when the first query is run against it.
Usually, postgresql server processes are spawned on a per-connection basis.
During the life time of a server process, the instance is cached. That means that if you have to keep references to resources such as connections, you should establish them in the init method and cache them as instance attributes._
I am not familiar with python. What exactly needs to be written in the _init method to have a fast first request?