src/test/regress/expected/create_extension_fail.out

--
-- Tests for return correct error from qe when create extension error
-- The issue: https://github.com/greenplum-db/gpdb/issues/11304
--
drop extension if exists gp_debug_numsegments;
NOTICE:  extension "gp_debug_numsegments" does not exist, skipping
create extension if not exists gp_inject_fault;
select gp_inject_fault('create_function_fail', 'error', dbid) from gp_segment_configuration where role = 'p' and content = 2;
 gp_inject_fault 
-----------------
 Success:
(1 row)

-- the following create exetension statement will fail due to fault-ibject
-- the failure happens during Create-UDFs for the extension on QEs, QDs
-- should catch the error and reset the global state variables in QEs correctly.
-- Previously, QD's error message is confusing, now it should print the direct
-- failing reason from QE instead of "query plan with multiple segworker groups is not supported".
create extension gp_debug_numsegments;
ERROR:  fault triggered, fault name:'create_function_fail' fault type:'error'  (seg2 127.0.1.1:7004 pid=6624)
-- Next we need to test the global state vars of extension creating
-- is reset correctly in the above error handling. Since we do not
-- destroy gangs, the next create table statement will reuse the
-- write gang of the above create extension statement. If the
-- global vars (creating_extension, CurrentExtensionObject) are
-- not resetting, then the following create statement when executing
-- on QEs will consider itself need to bind the table to the
-- extension and add wrong dependency with a non-exist extension (the
-- above extension creation statement aborts). Finally we can never
-- drop the table. The cases verify this by showing we can still
-- drop the table from a new session. A new session to drop is critical
-- for this test since new session have a clean environment.
create table t_11304(a int);
NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'a' as the Apache Cloudberry data distribution key for this table.
HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
-- start a new session so that make sure a new gang
\c
drop table t_11304;
select gp_inject_fault('create_function_fail', 'reset', dbid) from gp_segment_configuration where role = 'p' and content = 2;
 gp_inject_fault 
-----------------
 Success:
(1 row)

--
-- Another Test from Issue: https://github.com/greenplum-db/gpdb/issues/12713
-- Similar comments please refer above case's
--
create role user_12713;
NOTICE:  resource queue required -- using default resource queue "pg_default"
set role user_12713;
-- the will fail due to privilege
create extension gp_debug_numsegments;
ERROR:  permission denied to create extension "gp_debug_numsegments"
HINT:  Must be superuser to create this extension.
create table t_12713(a int);
NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'a' as the Apache Cloudberry data distribution key for this table.
HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
\c
drop table t_12713;
drop role user_12713;
--
-- Another Test from Issue: https://github.com/greenplum-db/gpdb/issues/6716
--
drop extension if exists gp_inject_fault;
create schema issue6716;
create extension gp_inject_fault with schema issue6716;
drop extension gp_inject_fault;
drop schema issue6716;