Re: Help with postgresql memory issue

The machine is running a moderate load. This is running on a Solaris
Zone.

Top is showing:

load averages: 2.49, 4.00, 3.78; up
124
+
12
:
24
:
47
16
:04:21
46 processes: 45 sleeping, 1 on cpu
CPU states: 76.6% idle, 14.6% user, 8.8% kernel, 0.0% iowait, 0.0%
swap
Memory: 32G phys mem, 942M free mem, 76G swap, 74G free swap

PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
5069 postgres 1 52 0 167M 20M sleep 0:04 13.50% postgres
902 postgres 1 1 0 167M 21M sleep 0:12 6.39% postgres
5068 postgres 1 59 0 167M 21M sleep 0:01 4.92% postgres
5070 postgres 1 59 0 166M 20M sleep 0:00 3.72% postgres
27817 postgres 1 59 0 167M 22M sleep 0:23 1.43% postgres
903 postgres 1 59 0 157M 11M sleep 0:02 1.14% postgres
23594 postgres 1 59 0 148M 2096K sleep 0:10 0.11% postgres
5510 brooks 1 59 0 5624K 2184K cpu 0:00 0.10% top
23598 postgres 1 59 0 6404K 1680K sleep 0:11 0.10% postgres
23595 postgres 1 59 0 148M 1852K sleep 0:01 0.01% postgres
23597 postgres 1 59 0 6220K 1556K sleep 0:00 0.01% postgres
24870 root 30 39 0 7060K 3332K sleep 7:01 0.00% nscd
736 brooks 1 59 0 6292K 2060K sleep 0:00 0.00% sshd
23596 postgres 1 59 0 148M 2024K sleep 0:00 0.00% postgres
24828 root 13 29 0 9300K 2128K sleep 2:02 0.00% svc.st

And vmstat shows:

kthr memory page disk
faults cpu
r b w swap free re mf pi po fr de sr s0 s1 s2 s3 in sy cs
us sy id
0 0 0 74805352 2910024 373 4154 96 7 7 0 0 -0 52 19 19 4561 230770
6889 11 13 76

On 28-Oct-09, at 4:01 PM, Tom Lane wrote:

> Brooks Lyrette <brooks(dot)lyrette(at)gmail(dot)com> writes:
>> I'm new to postgres and it seems my server is unable to fork new
>> connections.
>
>> LOG: could not fork new process for connection: Not enough space
>
> For what I suppose is a lightly loaded machine, that is just plain
> weird. What's the platform exactly? Is it possible that the
> postmaster
> is being launched under very restrictive ulimit settings?
>
> If it's a Unix-ish machine, it would be useful to look at "top" and
> "vmstat" output to see if the machine is under severe memory
> pressure for some reason.
>
> regards, tom lane