Tidb缩减tikv机器

生产环境下，如何缩减机器？

1、首先是检查出来那个tikv节点需要缩减

[tidb@:vg_adn_tidbCkhsTest:172.31.30.62 ~/tidb-ansible]$/home/tidb/tidb-ansible/resources/bin/pd-ctl -u "http://172.31.30.62:2379" -d store 
{
  "count": 5,
  "stores": [
    {
      "store": {
        "id": 1,
        "address": "172.31.17.203:20171",
        "labels": [
          {
            "key": "host",
            "value": "tikv3"
          }
        ],
        "version": "2.1.0-rc.5",
        "state_name": "Up"
      },
      "status": {
        "capacity": "2.0 TiB",
        "available": "1.9 TiB",
        "leader_count": 297,
        "leader_weight": 1,
        "leader_score": 25792,
        "leader_size": 25792,
        "region_count": 1533,
        "region_weight": 1,
        "region_score": 127200,
        "region_size": 127200,
        "start_ts": "2019-01-02T12:56:00+08:00",
        "last_heartbeat_ts": "2019-01-02T14:42:25.586838854+08:00",
        "uptime": "1h46m25.586838854s"
      }
    },
        "version": "2.1.0-rc.5",
        "state_name": "Up"
      },
      "status": {
        "capacity": "2.0 TiB",
        "available": "1.9 TiB",
        "leader_count": 344,
        "leader_weight": 1,
        "leader_score": 25450,
        "leader_size": 25450,
        "region_count": 807,
        "region_weight": 1,
        "region_score": 63203,
        "region_size": 63203,
        "start_ts": "2019-01-02T12:53:05+08:00",
        "last_heartbeat_ts": "2019-01-02T14:42:33.18874117+08:00",
        "uptime": "1h49m28.18874117s"
      }
    },
    {
      "store": {
        "id": 7,
        "address": "172.31.26.133:20172",
        "labels": [
          {
            "key": "host",
            "value": "tikv2"
          }
        ],
        "version": "2.1.0-rc.5",
        "state_name": "Up"
      },
      "status": {
        "capacity": "955 GiB",
        "available": "945 GiB",
        "leader_count": 313,
        "leader_weight": 1,
        "leader_score": 25532,
        "leader_size": 25532,
        "region_count": 726,
        "region_weight": 1,
        "region_score": 63997,
        "region_size": 63997,
        "start_ts": "2019-01-02T12:54:51+08:00",
        "last_heartbeat_ts": "2019-01-02T14:42:29.339673448+08:00",
        "uptime": "1h47m38.339673448s"
      }
    }
  ]
}

我想缩减id为7的机器：

2、找出id号后开始缩减，

[tidb@:vg_adn_tidbCkhsTest:172.31.30.62 ~/tidb-ansible]$/home/tidb/tidb-ansible/resources/bin/pd-ctl -u "http://172.31.30.62:2379" -d store delete 7
Success!

3、检测该机器是否已经真正下线，

[tidb@:vg_adn_tidbCkhsTest:172.31.30.62 ~/tidb-ansible]$/home/tidb/tidb-ansible/resources/bin/pd-ctl -u "http://172.31.30.62:2379" -d store 7
{
  "store": {
    "id": 7,
    "address": "172.31.26.133:20172",
    "state": 1,
    "labels": [
      {
        "key": "host",
        "value": "tikv2"
      }
    ],
    "version": "2.1.0-rc.5",
    "state_name": "Offline"
  },

注意：此处显示为“Offline”并不表示真正的下线，下线需要几分钟，因此我们过10分钟再来查看一下

4、再次检查机器

[tidb@:vg_adn_tidbCkhsTest:172.31.30.62 ~/tidb-ansible]$/home/tidb/tidb-ansible/resources/bin/pd-ctl -u "http://172.31.30.62:2379" -d store 7
{
        "key": "host",
        "value": "tikv2"
      }
    ],
    "version": "2.1.0-rc.5",
    "state_name": "Tombstone"

注意：只有当此处变成“Tombstone”才表示机器真正的下线，接下来我们可以继续进行下一步了

4、下线成功后，就可以单独关闭这个节点了

ansible-playbook stop.yml -l 172.16.10.9

5、更新inventory.ini文件，将这个机器注释掉。

6、更新grafana

ansible-playbook rolling_update_monitor.yml --tags=prometheus