Bug #6287
Updated by François ARMAND almost 10 years ago
I have 4 GB RAM configured. Postgres is configured (on the advice of ncharles) with 800M shared_buffers.
The output of top for mem usage is:
<pre># top -b -n 1 | grep -i postgres
5870 postgres 20 0 953372 27496 27232 S 0.0 0.7 1:22.79 postgres
5872 postgres 20 0 954036 796240 795264 S 0.0 19.7 16:33.60 postgres
5873 postgres 20 0 953776 755616 754952 S 0.0 18.7 4:00.80 postgres
5874 postgres 20 0 953776 8892 8660 S 0.0 0.2 3:36.56 postgres
5875 postgres 20 0 954596 1480 992 S 0.0 0.0 0:54.97 postgres
5876 postgres 20 0 102420 836 360 S 0.0 0.0 5:55.41 postgres
6687 postgres 20 0 954968 801124 799644 S 0.0 19.8 175:37.15 postgres
18559 postgres 20 0 956636 816872 813944 S 0.0 20.2 16:33.29 postgres
18562 postgres 20 0 957768 828240 823896 S 0.0 20.5 19:53.75 postgres
18563 postgres 20 0 957732 791764 787456 S 0.0 19.6 8:28.75 postgres
18615 postgres 20 0 961124 817096 810044 S 0.0 20.2 9:46.15 postgres
18616 postgres 20 0 960116 828488 822224 S 0.0 20.5 25:43.11 postgres
# top -b -n 1 | grep -i slapd
5834 root 20 0 2669604 292584 17892 S 0.0 7.2 32:43.12 slapd
# top -b -n 1 | grep -i java
18190 root 20 0 2596388 1.226g 5048 S 0.0 31.8 242:36.77 java
# top -b -n 1 | grep -i rsyslog
1074 syslog 20 0 769460 7280 1924 S 31.7 0.2 160:58.41 rsyslogd
# top -b -n 1 | grep -i cf-serverd
2583 root 20 0 310312 9460 1532 S 0.0 0.2 6:35.29 cf-serverd</pre>
Edit (FAR): for more context, the installation is small, with around 100 nodes managed.