|
|
high_sql_data-1570447913732.txt (14,137 bytes) high_sql1.JPG (59,282 bytes) high_sql2.JPG (65,087 bytes) |
|
|
> Why pgpool so frequently query this statements and how to minimize it ? When Pgpool-II receives a client query, Pgpool-II need to inquiry PostgreSQL's system catalogs to determine that - if a table in client's query is a temporary table or not - if a table in client's query is an unlogged table or not - if a function in client's query is "immutable" or not Because such client's queries should be sent to PostgreSQL primary server only. So, Pgpool-II issues queries to PostgreSQL's system catalogs, when a table or function appears in client's query. However, this behaviour is improved in the coming major release Pgpool-II 4.1. In 4.1 new parameter "enable_shared_relcache" is added. If enable_shared_relcache is on, the relation stored in memory cache can be shared among Pgpool-II child processes. This feature can reduce the queries issued to PostgreSQL's system catalogs. |
|
|
Thx, pengbo |
- Anonymous
