On Fri, 3 Oct 2003, Oleg Bartunov wrote:
> Hmm,
>
> it's weird. Could you provide us with backtrace ?
Bear in mind the last patch should have fixed the specific point of failure but
something, possibly tsearch2 possibly something else, was obviously pushing the
process's memory requirements upto/beyond what the kernel was prepared or able
to assign to it.
> On Tue, 30 Sep 2003 psql-mail@freeuk.com wrote:
>
> > After applying the patches supplied so far and also trying the lastest
> > stable tar.gz for tsearch2 ( downloaded 24th of september)
> >
> > I am still experiencing the same issue as previously described:
> > I try to do a
> >
> > SELECT to_tsvector( 'default', 'some text' )
> >
> > The backend crashes.
> >
> > SELECT to_tsvector( 'default', <string of whitespace> )
> > does not crash
> >
> > Any more advice or suggestions?
As Oleg says, another back trace would be good. Perhaps the postgres.conf
settings for memory and the resource limits applied to the backend by the
system.
A small test case would help, I think you have changed from the default
dictionary for example.
I get:
=> select to_tsvector('default', 'some text');
to_tsvector
-------------
'text':2
(1 row)
for the default setup.
--
Nigel Andrews
