On Wed, Feb 09, 2011 at 02:15:06PM -0500, Tom Lane wrote:
- David Kerr <dmk@mr-paradox.net> writes:
- > Ok, I found the bad entries, 2 tables a sequence and 2 primary key indexes are associated
- > with the wrong (invalid / nonexistant ) schema.
-
- > However, there are correct entries for those objects as well. So these are definitly just dangling references.
-
- Please be more specific. What are the bad entries exactly (what values,
- in which columns of what catalogs) and what do you mean by "there are
- correct entries as well"?
pg_class has 5 relname entries associated with a relnamespace (58698) that doesn't exist in pg_namespace.
Those relname entries corrispond to 2 tables and their PKs and a sequence (for the PK of one of the tables).
Those objects do have valid entries in pg_class as well (they point to a different relnamespace)
Is that more clear?
- > When we do a data "refresh" here, i do a drop <schema> cascade; in the DB and then
- > pg_restore -Fc --disable-triggers
-
- Hm. We have seen occasional reports of drop cascade failing to delete
- all the dependent objects, but it's pretty hard to see how that could
- happen ...
I agree!
Dave
