CODEWITHSUNDEEP
elasticsearchlogstashkibanaelastic-stack
Nayanajith
Nayanajith

Reputation: 41

Elastic search cluster is shown as red. How to recover?

I have a ES cluster showing the status as red.

{ "cluster_name": "elasticsearch", "status": "red", "timed_out": false, "number_of_nodes": 10, "number_of_data_nodes": 7, "active_primary_shards": 4431, "active_shards": 8862, "relocating_shards": 12, "initializing_shards": 0, "unassigned_shards": 20 }

I have so far tried to rolling restart the data nodes after disabling shard allocation, but cannot see any progress. Could someone be able to provide me a few hints on what to check further?

Upvotes: 4

Views: 14241

Answers (2)

RicHincapie
RicHincapie

Reputation: 3973

As this post states, you may try to run this at your elasticsearch master so it tries to realocate the shred:

curl -XPOST localhost:9200/_cluster/reroute?retry_failed

In my case, it was able to recover the data and get the following message:

Elasticsearch cluster is green. Shards: 108 active, 0 initializing, 0 relocating, 0 unassigned...

Upvotes: 1

Marko Vranjkovic
Marko Vranjkovic

Reputation: 6869

You have "unassigned_shards": 20 in health check response. This is most likely caused because data nodes in the Elasticsearch cluster lack free storage space. As explained here you can:

  • Use the /_cat/indices Elasticsearch API to determine which of the indices are unassigned to nodes in your cluster

  • You can also use the _cat/allocation?v API to check shard allocation and disk usage.

Upvotes: 4

Related Questions