Endless creating index after delete another index

Elastic Stack Elasticsearch
5

Solved by allocating all the failed shards instead of deleting them.

POST /_cluster/reroute
{
  "commands": [
    {
      "allocate_empty_primary": {
        "index": "nginx_log-2018-06-20",
        "shard": 0,
        "node": "uL9qINwtQqmMpSEnugloEg",
        "accept_data_loss": true
      }
    }
  ]
}

After allocated these indices , you can delete them.

Note: this operation may cause data loss.