goroutine leak: goroutine increase during the long-term operation

[bufferflies]

Bug Report

The count of pd goroutine continue increase in long-term cluster

What did you do?

Trigger condition: remove scheduler

#/bin/bash

for i in $(seq 1 50); do
tiup ctl:v6.5.6 pd -u  http://127.0.0.1:2379 scheduler remove balance-leader-scheduler
sleep 10
tiup ctl:v6.5.6 pd -u  http://127.0.0.1:2379 scheduler add balance-leader-scheduler
sleep 20
done

root cause:

pd/server/cluster/coordinator.go

Lines 696 to 700 in 77d6f5b

	decoder := schedule.ConfigSliceDecoder(schedulerCfg.Type, schedulerCfg.Args)
	tmp, err := schedule.CreateScheduler(schedulerCfg.Type, schedule.NewOperatorController(c.ctx, nil, nil), storage.NewStorageWithMemoryBackend(), decoder)
	if err != nil {
	return err
	}

What did you expect to see?

The count of goroutine is steady

What did you see instead?

What version of PD are you using (pd-server -V)?

v6.5.6