The following bug has been logged on the website:
Bug reference: 12832
Logged by: Marko Tiikkaja
Email address: marko@joh.to
PostgreSQL version: 9.3.6
Operating system: Linux
Description:
Hi,
We've occasionally been having a problem where a script does roughly this:
psql -d dbname -c "CREATE TEMPORARY FUNCTION pg_temp.foo() AS ..; <run a
query using the temporary function a number of times>"
pg_dump -d dbname
And the pg_dump call ERRORs out like this:
pg_dump: [archiver (db)] query failed: ERROR: cache lookup failed for
function 2540046
pg_dump: [archiver (db)] query was: SELECT tableoid, oid, proname, prolang,
pronargs, proargtypes, prorettype, proacl, pronamespace,
pg_catalog.pg_get_function_identity_arguments(oid) AS proiargs,(SELECT
rolname FROM pg_catalog.pg_roles WHERE oid = proowner) AS rolname FROM
pg_proc p WHERE NOT proisagg AND (pronamespace != (SELECT oid FROM
pg_namespace WHERE nspname = 'pg_catalog')
AND NOT EXISTS (SELECT 1 FROM pg_depend WHERE classid =
'pg_proc'::regclass AND objid = p.oid AND deptype = 'i'))
The problem seems to be that the snapshot of pg_proc the pg_dump session has
sees the function, but by the time it gets to run
pg_get_function_identity_arguments() on it, it's been cleaned up by the
backend which created it. Now this is an issue with normal functions as
well if they're dropped concurrently with a pg_dump run, but running into
problems like this when doing everything sequentially seems more astonishing
to my little brain.
Any thoughts on adding NOT pg_is_other_temp_schema(pg_proc.pronamespace)
quals to the queries which look at pg_proc to counter this problem?
Obviously this can be worked around in client code with a strategically
placed DISCARD TEMP, but in general this seems like a POLA violation.
