High consumns memory - Mailing list pgsql-admin

From Anderson Valadares
Subject High consumns memory
Date
Msg-id fa11ab140904141105h496a9cbasf2dcebcf5b5a46bd@mail.gmail.com
Whole thread Raw
List pgsql-admin
I have a software developed in Delphi as a Windows Service, but, i don't know why, it consumns an unexpected large system memory (1.3g).
The service access PostgresSQL by ODBC driver (psqlodbc_08_03_0400) and it consist simply of a loop calling a procedure PL/PGSQL. How to discover what is causing or why this high memory usage ? What objects are being used on this session ?
Software developed in Delphi 7 as a windows service.
PostgresSQL 8.3.6 Database with PostGis extension

Server p52a
S.O.: Red Hat Enterprise Linux AS release 4 (Nahant Update 1)
      linux 2.6.9-11.EL #1 SMP ppc64 ppc64 ppc64 GNU/Linux
S.O. information

Date 13/04/2009

top - 15:20:35 up 10 days, 14:22,  1 user,  load average: 0.80, 0.66, 0.63

Tasks: 134 total,   2 running, 132 sleeping,   0 stopped,   0 zombie
Cpu(s):  3.2% us,  0.1% sy,  0.0% ni, 96.2% id,  0.3% wa,  0.1% hi,  0.1% si
Mem:   4107392k total,  1552168k used,  2555224k free,    10172k buffers
Swap:  2031608k total,    79740k used,  1951868k free,  1402644k cached
 
  PID USER      PR  NI  VIRT SWAP  RES  SHR DATA CODE S %CPU %MEM    TIME+  COMMAND
 1843 postgres  16   0  857m 664m 193m 186m 7120 3540 R 19.9  4.8   0:06.13 postgres: citgis citgis 10.255.100.73(1557) SELECT
 

Date 14/04/2009

top - 12:51:44 up 11 days, 11:54,  1 user,  load average: 0.41, 0.45, 0.42
Tasks: 128 total,   1 running, 127 sleeping,   0 stopped,   0 zombie
Cpu(s):  3.8% us,  0.1% sy,  0.0% ni, 95.7% id,  0.2% wa,  0.1% hi,  0.0% si
Mem:   4107392k total,  4097504k used,     9888k free,    48252k buffers
Swap:  2031608k total,    62680k used,  1968928k free,  2804800k cached

  PID USER      PR  NI  VIRT SWAP  RES  SHR DATA CODE S %CPU %MEM    TIME+  COMMAND
 1843 postgres  15   0 1896m 114m 1.7g 737m 1.0g 3540 S 26.9 44.4 181:59.47 postgres: citgis citgis 10.255.100.73(1557) idle

As shown in column DATA(7120) on 03/04 and on 04/04 (1.0g) been a significant increase in the consumption of memory.


postgresql.conf information:
 
               name               |                        setting       | unit
 ---------------------------------+--------------------------------------+------
  archive_mode                    | on                                   |
  autovacuum_analyze_scale_factor | 0.4                                  |
  autovacuum_analyze_threshold    | 500                                  |
  autovacuum_vacuum_threshold     | 1000                                 |
  checkpoint_segments             | 15                                   |
  checkpoint_timeout              | 1800                                 | s
  default_statistics_target       | 50                                   |
  effective_cache_size            | 249600                               | 8kB
  fsync                           | on                                   |
  logging_collector               | on                                   |
  maintenance_work_mem            | 409600                               | kB
  max_connections                 | 100                                  |
  max_fsm_pages                   | 3458000                              |
  shared_buffers                  | 64000                                | 8kB
  wal_buffers                     | 100                                  | 8kB
  work_mem                        | 5120                                 | kB

pgsql-admin by date:

Previous
From: "Kevin Grittner"
Date:
Subject: Re: pg error log file monitor
Next
From: Benjamin Minshall
Date:
Subject: Backup only changed databases