On 04/03/03 20:21, Tom Lane wrote:
> Lonni Friedman <lfriedman@vasoftware.com> writes:
>> however, if i attempt to select just bin_data (which holds a 6MB M$-Word
>> doc) it attempts to dump it to screen for a few hundred lines, and then
>> segfaults.
>
> Wait a second. Which side is segfaulting here? I've been assuming that
> you were talking about a backend segfault, but maybe I misunderstood
> completely. If it's psql dumping core then we have quite a different
> issue.
its psql, not the backend.
>> i'm honestly not sure how to do this. the segfault is not producing a
>> core file that i can find. perhaps i'd need to tinker with ulimit?
>
> Probably. "ulimit -c 0" is a common default these days. Try "ulimit -c
> unlimited" before starting the postmaster (resp. psql).
ok, is this even neccesary if the problem with the segfault is psql, and
not the backend?