Evict leader does't work as expected

[weekface]

    {
      "store": {
        "id": 1,
        "address": "e2e-cluster1-tikv-0.e2e-cluster1-tikv-peer.e2e-cluster1.svc:20160",
        "labels": [
          {
            "key": "host",
            "value": "kube-node-1"
          },
          {
            "key": "rack",
            "value": "rack1"
          }
        ],
        "version": "3.0.0-rc.1",
        "state_name": "Up"
      },
      "status": {
        "capacity": "252 GiB",
        "available": "119 GiB",
        "leader_weight": 1,
        "region_count": 19,
        "region_weight": 1,
        "region_score": 19,
        "region_size": 19,
        "start_ts": "2019-06-21T08:45:53Z",
        "last_heartbeat_ts": "2019-06-21T10:29:44.570183932Z",
        "uptime": "1h43m51.570183932s"
      }
    }
  ]
}
[jenkins@k8s-e2e-test ~]$ curl 127.0.0.1:2379/pd/api/v1/schedulers
Handling connection for 2379
[
  "evict-leader-scheduler-4",
  "evict-leader-scheduler-1",
  "balance-region-scheduler",
  "balance-leader-scheduler",
  "balance-hot-region-scheduler",
  "label-scheduler",
  "evict-leader-scheduler-5"
]

There three evict-leader-scheduler-* all exist in the schedulers api. This is a bug.

[weekface]

failed again:

I0621 19:13:39.247192       1 pods.go:152] savely delete pod namespace ns1 name cluster1-pd-2 leader name cluster1-pd-0
I0621 19:13:40.847572       1 actions.go:669] checking tidb cluster [ns1/cluster1] status
I0621 19:13:56.573789       1 pods.go:152] savely delete pod namespace ns1 name cluster1-pd-1 leader name cluster1-pd-0
I0621 19:14:22.732340       1 pods.go:152] savely delete pod namespace ns1 name cluster1-pd-0 leader name cluster1-pd-2
E0621 19:14:45.961808       1 actions.go:1475] check [ns1/cluster1] TiDB instance [0] configuration updated failed: desired [1024], actual [1000] not equal
E0621 19:15:45.952541       1 actions.go:1475] check [ns1/cluster1] TiDB instance [0] configuration updated failed: desired [1024], actual [1000] not equal
E0621 19:16:40.704683       1 pods.go:188] failed to evict leader from ns1/cluster1-tikv-2, before: 1, now: 29
E0621 19:16:45.978560       1 actions.go:1475] check [ns1/cluster1] TiDB instance [0] configuration updated failed: desired [1024], actual [1000] not equal

    {
      "store": {
        "id": 5,
        "address": "cluster1-tikv-2.cluster1-tikv-peer.ns1.svc:20160",
        "labels": [
          {
            "key": "host",
            "value": "172.16.4.149"
          },
          {
            "key": "rack",
            "value": "rack0"
          }
        ],
        "version": "3.0.0-rc.2",
        "state_name": "Up"
      },
      "status": {
        "capacity": "1.4 TiB",
        "available": "1.4 TiB",
        "leader_count": 9,
        "leader_weight": 1,
        "leader_score": 292,
        "leader_size": 292,
        "region_count": 29,
        "region_weight": 1,
        "region_score": 346,
        "region_size": 346,
        "start_ts": "2019-06-21T19:17:46Z",
        "last_heartbeat_ts": "2019-06-22T00:41:28.398805448Z",
        "uptime": "5h23m42.398805448s"
      }
    }
  ]
}
[root@172.16.4.149 ~]# curl 10.233.29.214:2379/pd/api/v1/schedulers
[
  "label-scheduler",
  "balance-region-scheduler",
  "balance-leader-scheduler",
  "balance-hot-region-scheduler"
]

@xiaojingchen @shuijing198799 PTAL