huge memory of Postgresql backend process

From: "James Pang (chaolpan)" <chaolpan(at)cisco(dot)com>
To: PostgreSQL mailing lists <pgsql-bugs(at)lists(dot)postgresql(dot)org>
Subject: huge memory of Postgresql backend process
Date: 2022-09-08 06:10:28
Message-ID: PH0PR11MB51911F681F37F55A92DA4C01D6409@PH0PR11MB5191.namprd11.prod.outlook.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Hi,
We just switch from Oracle to PGV13, all of these tables got partition (256 HASH partition) , when UPDATE/DELETE ... WHERE . The memory increase verify quickly until >130M. not too much data in these tables. When we switch to PGV14.5 , test again, still see similar memory consumption issue. For example , one Postgresql JDBC client ,

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ nMaj nMin WCHAN COMMAND
551701 postgres 20 0 45.6g 131468 11664 S 0.0 0.1 0:03.11 0 35k do_epoll_w+ postgres:

When we rebuild the table to no-partitioned tables , same table columns and same data size, the "RES" memory only 25MB. For both partitioned and non-partitioned table, the UPDATE/DELETE run fast. The only difference is big difference on memory consumption. And even the session is IDLE there long time, the backend process did NOT release memory to OS. Is it expected to see that ?

Server memory 128G, shared_buffer 45G, effective_cache_size=80G. PG server both V13.4 and V14.5 has same issue. Postgresql JDBC client driver 42.3.4

Thanks,

James

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message David Rowley 2022-09-08 08:36:04 Re: Postgres offset and limit bug
Previous Message David G. Johnston 2022-09-08 06:04:45 A problem about LEAKPROOF of postgresql