Sounds like you are running of filedescriptors That can be due to
following factors:
1. Kernel limit is exhausted (not likely, i think default is
1024fds/process), and needs to be adjusted
2. You are exceeding your ulimit for fds. type ulimit -a and see what's it
set to.
3. Postgres is leaking FDs...Not likely. But cd /proc/<pid>/fds and see
how many of them are used. If the number is steadily growing with time,
its likely to be the case.
-alex
On Tue, 14 Mar 2000, Uwe Scholz wrote:
> Hallo,
> I get this message on each connect (the logfilesize is 1M in 3 days).
> I have a 24 hour production datebase with alltogether over 600,000 rows since one
>year. I never had trouble until I changed the machine an the postgres version.
> Before I had a celeron 300 and postgreSQL 6.4.2.
> Now I use a dual P3-500 and postgreSQL 6.5.3, SuSE-Linux 6.3, kernel 2.2.14(SMP),
>glibc 2.1.2, egcs-2.91.66
> And the badest is: the server crashes after 2 days with AllocateFile: too many
>private FDs demanded.
> What can I do ? I have only sleepless nights since them.
>
> Uwe Scholz
>
> sorry for my bad english
>
>
>
> ---
> Microsoft is not the Answer -
> Microsoft is the Question,
> and the Answer is no !
>
>
>
