"Johnson, Shaunn" <SJohnson6@bcbsm.com> writes:
> I can't find anything in the serverlog file that suggest that there are
> errors occurring - or what exactly is failing. The only thing in the
> /var/log/messages is whenever someone kicks off psql for one reason
> or another ... but that doesn't help.
In that case you should first work on figuring out why your logging
setup doesn't work. A backend crash should definitely provoke all kinds
of log output --- not only a "child process exited with status NN"
message, but also complaints from any concurrent backends that got told
to quit.
The simplest, most reliable setup is to not use syslog logging but just
direct the postmaster's stdout/stderr into some file. Try that, and be
sure not to use -S in the postmaster options.
> I imagine that maybe the
> table is corrupt, but how can I tell?
Sounds that way to me too. If you just do a "select * from table" do
you get a similar crash?
regards, tom lane
