joseph speigle <joe(dot)speigle(at)jklh(dot)us> writes:
> this simple update closes the connection.
My first guess is there's some data corruption in the table or one of
its indexes. It would be useful to get a stack trace showing where
in the backend the core dump occurs --- can you manage that?
Other things to check: can you do a "select * from calling" without
crashing? What shows up in the postmaster log when the crash occurs?
If you just want to fix the database ASAP, you could try a REINDEX
on the table --- if the problem is in the indexes that should get
rid of it. But it would also destroy the evidence of exactly what
went wrong.
regards, tom lane