Bug #540: pg_dump a simple table can fail if other view/triggers are faulty

On Sun, 16 Dec 2001 pgsql-bugs@postgresql.org wrote:

luby liao (liao@sandiego.edu) reports a bug with a severity of 2
The lower the number the more severe it is.

Short Description
pg_dump a simple table can fail if other view/triggers are faulty

Long Description
We use Postgresql 7.1.2 on a Redhat 7.0 Linux PC. My CS145 students
all work in a database called cs145. Some students introduced
unintended faulty views and triggers into the database. When a
student tried to pg_dump his own tables, he got error messages that
say that other people's views or triggers were faulty.

We're going to need some additional information for this to be a
complete report. What options were used to pg_dump, how far the dump
got and what messages were reported.
It'd be nice to know what was meant by faulty views and triggers or
an example to reproduce the problem.

Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:

We're going to need some additional information for this to be a
complete report.

It doesn't really require investigation. pg_dump pulls the entire
schema before deciding what parts of it need to be dumped; therefore
schema-level problems are likely to cause pg_dump to error out even
if that part of the schema theoretically didn't need to be examined.

I am disinclined to invest any effort in changing this, in part because
I doubt it could be changed without making it impossible to do other
things that we want to do in the near future (cf. nearby discussion
about dumping sequences associated with the target table).

regards, tom lane

Tom and Stephan, thank you so much for the prompt reply. cheers, Luby