[Bug] unstable resource group cases: resgroup_cancel_terminate_concurrency/resgroup_dumpinfo/resgroup_views

[avamingli]

Cloudberry Database version

main

What happened

https://github.com/cloudberrydb/cloudberrydb/actions/runs/11102859132/job/30844343864

DIFF FILE: ../gpdb_src/src/test/isolation2/regression.diffs
----------------------------------------------------------------------

diff -I HINT: -I CONTEXT: -I GP_IGNORE: -U3 /code/gpdb_src/src/test/isolation2/expected/resgroup/resgroup_cancel_terminate_concurrency.out /code/gpdb_src/src/test/isolation2/results/resgroup/resgroup_cancel_terminate_concurrency.out
--- /code/gpdb_src/src/test/isolation2/expected/resgroup/resgroup_cancel_terminate_concurrency.out	2024-09-30 17:27:33.490609064 +0800
+++ /code/gpdb_src/src/test/isolation2/results/resgroup/resgroup_cancel_terminate_concurrency.out	2024-09-30 17:27:33.497609032 +0800
@@ -293,9 +293,10 @@
 1q: ... <quitting>
 2q: ... <quitting>
 DROP ROLE role_concurrency_test;
-DROP
+DETAIL:  owner of table pg_temp_27.tmp
+ERROR:  role "role_concurrency_test" cannot be dropped because some objects depend on it
 DROP RESOURCE GROUP rg_concurrency_test;
-DROP
+ERROR:  resource group is used by at least one role
 
 DROP VIEW rg_concurrency_view;
 DROP
diff -I HINT: -I CONTEXT: -I GP_IGNORE: -U3 /code/gpdb_src/src/test/isolation2/expected/resgroup/resgroup_dumpinfo.out /code/gpdb_src/src/test/isolation2/results/resgroup/resgroup_dumpinfo.out
--- /code/gpdb_src/src/test/isolation2/expected/resgroup/resgroup_dumpinfo.out	2024-09-30 17:28:21.174385431 +0800
+++ /code/gpdb_src/src/test/isolation2/results/resgroup/resgroup_dumpinfo.out	2024-09-30 17:28:21.176385421 +0800
@@ -43,7 +43,7 @@
 SELECT dump_test_check();
  dump_test_check 
 -----------------
- t               
+ f               
 (1 row)

What you think should happen instead

Pretty sure #649 has nothing to do with resource group.
It seems a unstable case of resource group.

How to reproduce

https://github.com/cloudberrydb/cloudberrydb/actions/runs/11102859132/job/30844343864

Operating System

CI OS

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.

[avamingli]

Failed case in another PR, memory_usage in resgroup/resgroup_views.out
https://github.com/cloudberrydb/cloudberrydb/actions/runs/11136689360/job/30949248752?pr=653

/code/gpdb_src/src/test/isolation2/results/resgroup/resgroup_views.out
--- /code/gpdb_src/src/test/isolation2/expected/resgroup/resgroup_views.out	2024-10-02 10:39:17.270264381 +0800
+++ /code/gpdb_src/src/test/isolation2/results/resgroup/resgroup_views.out	2024-10-02 10:39:17.271264371 +0800
@@ -14,7 +14,7 @@
 select groupname , groupid , cpu_usage , memory_usage from gp_toolkit.gp_resgroup_status_per_host s join gp_segment_configuration c on s.hostname=c.hostname and c.content=-1 and role='p' where groupname='default_group';
  groupname     | groupid | cpu_usage | memory_usage 
 ---------------+---------+-----------+--------------
- default_group | 6437    | 0.00      | 0.00         
+ default_group | 6437    | 0.00      | 0.01         
 (1 row)
 
 select * from gp_toolkit.gp_resgroup_role where rrrolname='gpadmin';


real	6m5.137s
user	0m21.717s
sys	0m4.744s

[jiaqizho]

memory usage is acceptable between 0.00 ~ 0.01 . so let we change the .sql/.out