集群恢复时卡住

版本：6.1.1
Nodes: 4(master+data)
          Indices: 3352
          Memory: 51GB / 123GB
         Total Shards: 33379
         Documents: 1,456,904,986
         Data: 492GB
 
国庆前进行了滚动重启，本以为数据会缓慢恢复，但是来了一看发现卡住了
查看一下任务

GET _cat/recovery?v&h=i,s,t,ty,st,fp,bp,shost,thost

 
查看线程

GET _cat/thread_pool?v

整个集群状态为yellow,只要是从其他节点恢复数据的任务全部卡住，这是我的集群设置

{

  "persistent": {

    "cluster": {

      "routing": {

        "allocation": {

          "allow_rebalance": "always",

          "cluster_concurrent_rebalance": "500",

          "enable": "all"

        }

      }

    },

    "discovery": {

      "zen": {

        "minimum_master_nodes": "2"

      }

    },

    "indices": {

      "recovery": {

        "max_bytes_per_sec": "600mb"

      }

    }

  },

  "transient": {

    "cluster": {

      "routing": {

        "allocation": {

          "node_concurrent_incoming_recoveries": "500",

          "node_initial_primaries_recoveries": "4",

          "enable": "all",

          "node_concurrent_outgoing_recoveries": "500"

        }

      }

    }

  }

}

请问这个问题出在哪。。。