Re: connections problem - Mailing list pgsql-hackers

From ivan
Subject Re: connections problem
Date
Msg-id Pine.LNX.4.56.0312262308160.28459@rex.anfa.pl
Whole thread Raw
In response to Re: connections problem  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers

from log :
LOG:  could not fork new process for connection: Resource temporarily unavailable
LOG:  could not fork checkpoint process: Resource temporarily unavailable
....


ulimit -a :

core file size        (blocks, -c) 0
data seg size         (kbytes, -d) unlimited
file size             (blocks, -f) unlimited
max locked memory     (kbytes, -l) unlimited
max memory size       (kbytes, -m) unlimited
open files                    (-n) 2048
pipe size          (512 bytes, -p) 8
stack size            (kbytes, -s) 15000
cpu time             (seconds, -t) unlimited
max user processes            (-u) 400
virtual memory        (kbytes, -v) unlimited

On Fri, 26 Dec 2003, Tom Lane wrote:

> ivan <iv@psycho.pl> writes:
> > the error is : could not fork new process for connection: Resource
> > temporarily unavailable
>
> We could have saved some time if you'd spelled out the full error
> message to begin with.
>
> This is undoubtedly a kernel-imposed restriction you are running into.
> I'm not sure if the failure is related to sheer number of processes,
> number of open files, or RAM/swap space, but one way or another you are
> trying to create more processes than your kernel will support.
>
> It seems moderately unlikely that twenty or thirty Postgres processes
> would by themselves be able to run the kernel out of resources (although
> if the problem is related to number of open files, reducing
> max_files_per_process might help).  What else do you have running on
> this machine?
>
>             regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>


pgsql-hackers by date:

Previous
From: Eric Ridge
Date:
Subject: Re: fulltext searching via a custom index type
Next
From: Tom Lane
Date:
Subject: Re: fulltext searching via a custom index type