Tom Lane wrote:
> =?iso-8859-1?q?Clodoaldo=20Pinto=20Neto?= <clodoaldo_pinto(at)yahoo(dot)com(dot)br> writes:
>> --- Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> escreveu:
>>>Look in the postmaster's log to see what happened.
>>LOG: server process (PID 9576) was terminated by signal 9
> Well, there's a smoking gun ...
> Assuming that you didn't deliberately kill -9 your server process,
> the only mechanism I've heard of that causes this is the Linux kernel's
> out-of-memory killer
In fact, looking at the original post, the evidence is there in the free
output - swap all used up, no shared/buffers left to speak of...
> $ free
> total used free shared buffers
> Mem: 759600 755712 3888 0 8380
> -/+ buffers/cache: 677188 82412
> Swap: 1539712 1539712 0
The sort_mem might be too high at 64MB - that's memory per sort and I
don't know how many simultaneous sorts you might have.
In response to
pgsql-general by date
|Next:||From: Francois Suter||Date: 2004-08-30 09:35:05|
|Subject: Technical recommendations|
|Previous:||From: Russell Smith||Date: 2004-08-30 06:38:49|
|Subject: timestamp and date behaviour with '-infinity'|