单节点 Elasticsearch5.6 健康状态为 yellow 问题的解决办法

单节点 Elasticsearch5.6 健康状态为 yellow 问题的解决办法

Administrator 1071 2020-10-27

项目环境中，单机部署的 Elasticsearch 刚部署完成时，health 状态为正常的 green，但是过了几天后，在查看 Elasticsearch 的 health，已经变为 yellow ：

[root@localhost ~]# curl '127.0.0.1:9200/_cluster/health?pretty'
{
  "cluster_name" : "master",
  "status" : "yellow",
  "timed_out" : false,
  "number_of_nodes" : 1,
  "number_of_data_nodes" : 1,
  "active_primary_shards" : 20,
  "active_shards" : 20,
  "relocating_shards" : 0,
  "initializing_shards" : 0,
  "unassigned_shards" : 20,
  "delayed_unassigned_shards" : 0,
  "number_of_pending_tasks" : 0,
  "number_of_in_flight_fetch" : 0,
  "task_max_waiting_in_queue_millis" : 0,
  "active_shards_percent_as_number" : 50.0
}

  • 根据上面的返回信息，集群的 unassigned_shards 为 20。

  • 单点部署的 Elasticsearch，默认分片的副本数为 1，而相同的分片不能在同一个节点上，所以就出现上面 unsigned shards 的问题。解决方法如下：

$ curl -X PUT "127.0.0.1:9200/_settings" -H 'Content-Type: application/json' -d'{"number_of_replicas":0}'

# 返回
{"acknowledged":true}
  • 再次查看 Elasticsearch health，已经转为 green：
[root@web-01 ~]# curl -X GET "127.0.0.1:9200/_cluster/health?pretty"
{
  "cluster_name" : "master",
  "status" : "green",
  "timed_out" : false,
  "number_of_nodes" : 1,
  "number_of_data_nodes" : 1,
  "active_primary_shards" : 20,
  "active_shards" : 20,
  "relocating_shards" : 0,
  "initializing_shards" : 0,
  "unassigned_shards" : 0,
  "delayed_unassigned_shards" : 0,
  "number_of_pending_tasks" : 0,
  "number_of_in_flight_fetch" : 0,
  "task_max_waiting_in_queue_millis" : 0,
  "active_shards_percent_as_number" : 100.0
}
ElasticSearch