Fix the issue of duplicate counting of num_executed in gp_toolkit.gp_resgroup_status

This is a defect of the original GPDB.
When enabling resource group management and the transaction switches from "Assign" to "Bypass" state, the "num_executed" counter is repeatedly counted.

Author: fairyfar

src/backend/utils/resgroup/resgroup.c

@@ -3700,7 +3700,6 @@ check_and_unassign_from_resgroup(PlannedStmt* stmt)
 	} while (!groupIncBypassedRef(&groupInfo));
 
 	bypassedGroup = groupInfo.group;
-	bypassedGroup->totalExecuted++;
 	pgstat_report_resgroup(bypassedGroup->groupId);
 	bypassedSlot.group = groupInfo.group;
 	bypassedSlot.groupId = groupInfo.groupId;

src/test/isolation2/expected/resgroup/resgroup_bypass.out

@@ -276,6 +276,24 @@ SELECT gp_inject_fault('func_init_plan_end', 'reset', 1);
 1q: ... <quitting>
 2q: ... <quitting>
 
+-- verify the increment of num_executed in gp_toolkit.gp_resgroup_status
+1: SET ROLE role_bypass;
+SET
+1: SELECT num_executed INTO temporary temp_num1 FROM gp_toolkit.gp_resgroup_status WHERE groupname='rg_bypass';
+SELECT 1
+1: SELECT num_executed INTO temporary temp_num2 FROM gp_toolkit.gp_resgroup_status WHERE groupname='rg_bypass';
+SELECT 1
+1: SELECT temp_num2.num_executed - temp_num1.num_executed AS delta FROM temp_num1, temp_num2;
+ delta 
+-------
+ 1     
+(1 row)
+1: DROP TABLE temp_num1;
+DROP
+1: DROP TABLE temp_num2;
+DROP
+1q: ... <quitting>
+
 -- cleanup
 -- start_ignore
 DROP TABLE t_bypass;

src/test/isolation2/sql/resgroup/resgroup_bypass.sql

@@ -133,6 +133,15 @@ SELECT gp_inject_fault('func_init_plan_end', 'reset', 1);
 1q:
 2q:
 
+-- verify the increment of num_executed in gp_toolkit.gp_resgroup_status
+1: SET ROLE role_bypass;
+1: SELECT num_executed INTO temporary temp_num1 FROM gp_toolkit.gp_resgroup_status WHERE groupname='rg_bypass';
+1: SELECT num_executed INTO temporary temp_num2 FROM gp_toolkit.gp_resgroup_status WHERE groupname='rg_bypass';
+1: SELECT temp_num2.num_executed - temp_num1.num_executed AS delta FROM temp_num1, temp_num2;
+1: DROP TABLE temp_num1;
+1: DROP TABLE temp_num2;
+1q:
+
 -- cleanup
 -- start_ignore
 DROP TABLE t_bypass;