[Bug] Some commands forgot to mark meta track calls unstable failed cases. #504
Description
Cloudberry Database version
PostgreSQL 14.4 (Cloudberry Database 1.5.4+dev.74.gc87fc6d8bb6 build dev) on x86_64-pc-linux-gnu, compi
led by gcc (Ubuntu 11.3.0-1ubuntu1~22.04.1) 11.3.0, 64-bit compiled on Jul 8 2024 13:50:22 (with assert
checking)
(1 row)What happened
Unstable case:
/code/gpdb_src/src/test/regress/results/aqumv.out
[15511](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15512)--- /code/gpdb_src/src/test/regress/expected/aqumv.out 2024-07-08 03:43:30.532971623 +0000
[15512](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15513)+++ /code/gpdb_src/src/test/regress/results/aqumv.out 2024-07-08 03:43:30.555971269 +0000
[15513](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15514)@@ -4,247 +4,132 @@
[15514](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15515) --
[15515](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15516) set optimizer = off;
[15516](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15517) create schema aqumv;
[15517](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15518)+DETAIL: Key (classid, objid, staactionname)=(2615, 24638, CREATE) already exists.
[15518](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15519)+ERROR: duplicate key value violates unique constraint "pg_statlastop_classid_objid_staactionname_index"
[15519](https://github.com/cloudberrydb/cloudberrydb/actions/runs/9832799441/job/27142431110?pr=501#step:5:15520) set search_path to aqumv;What you think should happen instead
Drop Schema command forget to drop meta track object.
While reading the codes, I find there are several places have the same issue.
- RemoveSchemaById
- RemovePublicationById
- RemovePolicyById
When some oids are used again, it will report error due to the unique index.
gpadmin=# \d+ pg_stat_last_operation;
Table "pg_catalog.pg_stat_last_operation"
Column | Type | Collation | Nullable | Default | Storage | Compression | St
ats target | Description
---------------+--------------------------+-----------+----------+---------+----------+-------------+---
-----------+-------------
classid | oid | | not null | | plain | |
|
objid | oid | | not null | | plain | |
|
staactionname | name | | not null | | plain | |
|
stasysid | oid | | not null | | plain | |
|
stausename | name | | not null | | plain | |
|
stasubtype | text | C | | | extended | |
|
statime | timestamp with time zone | | | | plain | |
|
Indexes:
"pg_statlastop_classid_objid_staactionname_index" UNIQUE CONSTRAINT, btree (classid, objid, staactio
nname)How to reproduce
gpadmin=# select * from pg_stat_last_operation;
classid | objid | staactionname | stasysid | stausename | stasubtype | statime
---------+-------+---------------+----------+------------+------------+---------
(0 rows)
gpadmin=# create schema xmen;
CREATE SCHEMA
gpadmin=# select * from pg_stat_last_operation;
classid | objid | staactionname | stasysid | stausename | stasubtype | statime
---------+-------+---------------+----------+------------+------------+-------------------------------
2615 | 16385 | CREATE | 10 | gpadmin | SCHEMA | 2024-07-08 13:57:18.716533+08
(1 row)
gpadmin=# drop schema xmen;
DROP SCHEMA
gpadmin=# select * from pg_stat_last_operation;
classid | objid | staactionname | stasysid | stausename | stasubtype | statime
---------+-------+---------------+----------+------------+------------+-------------------------------
2615 | 16385 | CREATE | 10 | gpadmin | SCHEMA | 2024-07-08 13:57:18.716533+08
(1 row)Operating System
Linux i-v3roa8c2 5.15.0-76-generic #83-Ubuntu SMP Thu Jun 15 19:16:32 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
Anything else
No response
Are you willing to submit PR?
- Yes, I am willing to submit a PR!
Code of Conduct
- I agree to follow this project's Code of Conduct.