Elasticsearch get rid of yellow status

Hi
I have Elasticsearch that I installed as a single node at the beginning and then the indices were created this configuration, after I added two other nodes and restarted elastic, and I always have the health status yellow, then I changed number_of_replicas set to 0 as I read on the internet and then put it back to 1

However even in this way the yellow status is still displayed; how I can get rid of this yellow status?

this is my elasticsearch config:

     cluster.name: cluster1
     node.name: cluster_node1
     network.host: 0.0.0.0
     http.port: 9200
     discovery.seed_hosts: ["env1", "env2", "env3"]
     cluster.initial_master_nodes: ["cluster_node1", "cluster_node2", "cluster_node3"]

And i have only this info in log :

    Cluster health status changed from [RED] to [YELLOW] (reason: [shards started [[.kibana_1] 
    [0]]]). 

no error o warning before this message.

Thanks.

I tried to delete kibana_1 index as stated in the previuos elastic log:

(Cluster health status changed from [RED] to [YELLOW] (reason: [shards started [[.kibana_1] ) :

so i ran:

DELETE /.kibana_1

and i received a success message.

Then i have green status in elastic but kibana dont work anymore, i see in the kibana log:

{"type":"log","@timestamp":"2021-01-14T10:38:19Z","tags":["info","savedobjects- 
service"],"pid":4704,"message":"Waiting until all Elasticsearch nodes are compatible with Kibana 
before starting saved objects migrations..."}
{"type":"log","@timestamp":"2021-01-14T10:38:19Z","tags":["info","savedobjects- 
service"],"pid":4704,"message":"Starting saved objects migrations"}
{"type":"log","@timestamp":"2021-01-14T10:38:49Z","tags":["warning","savedobjects- 
service"],"pid":4704,"message":"Unable to connect to Elasticsearch. Error: Request Timeout 
after 30000ms"}

Can anyone help me?
Thanks

Probably best not to go deleting random indices, especially ones managed by the stack. It will only break things.

A yellow cluster health means one or more shards are unassigned. If you want to understand why, use the cluster allocation explain API.

1 Like

Thanks for reply, i have any chance to recovery connection between kibana and elastic without reinstall all?

the connection between the two, seems to have broken after the deletion of the indexes

this is my kibana.yml file:

     server.port: 5601
     server.host: [ipaddress]
     server.name: "servername"
     elasticsearch.hosts:  [ipnode1, ipnode2, ipnode3]

and this is the log file of kibana:

    {"type":"log","@timestamp":"2021-01-15T14:26:14Z","tags":["warning","plugins- 
    discovery"],"pid":3176,"message":"Expect plugin \"id\" in camelCase, but found: 
    apm_oss"}
    {"type":"log","@timestamp":"2021-01-15T14:26:17Z","tags":["info","plugins- 
    service"],"pid":3176,"message":"Plugin \"visTypeXy\" is disabled."}
    {"type":"log","@timestamp":"2021-01-15T14:26:18Z","tags":["warning","legacy- 
    service"],"pid":3176,"message":"Some installed third party plugin(s) [apm_oss, 
    console_legacy, elasticsearch, kibana, newsfeed, region_map, status_page, 
    tile_map, timelion, ui_metric] are using the legacy plugin format and will no 
    longer work in a future Kibana release. Please refer to https://ela.st/kibana- 
    breaking-changes-8-0 for a list of breaking changes and https://ela.st/kibana- 
    platform-migration for documentation on how to migrate legacy plugins."}
   {"type":"log","@timestamp":"2021-01-15T14:26:18Z","tags":["info","plugins- 
   system"],"pid":3176,"message":"Setting up [42] plugins: 
   [usageCollection,telemetryCollectionManager,telemetry,kibanaLegacy,devTools,ui 
 Actions,statusPage,share,newsfeed,mapsLegacy,kibanaUtils,kibanaReact,inspector,embeddable,indexPatternManagement,esUiShared,discover,charts,bfetch,expressions,data,home,console,apm_oss,management,advancedSettings,telemetryManagementSection,visualizations,visTypeVislib,visTypeVega,visTypeTimeseries,visTypeTi 
 melion,visTypeTagcloud,visTypeTable,visTypeMetric,visTypeMarkdown,inputContr 
   olVis,savedObjects,navigation,visualize,dashboard,savedObjectsManagement]"}
   {"type":"log","@timestamp":"2021-01-15T14:26:19Z","tags":["info","savedobjects- 
   service"],"pid":3176,"message":"Waiting until all Elasticsearch nodes are 
   compatible with Kibana before starting saved objects migrations..."}
   {"type":"log","@timestamp":"2021-01-15T14:26:24Z","tags":["info","savedobjects- 
   service"],"pid":3176,"message":"Starting saved objects migrations"}
   {"type":"log","@timestamp":"2021-01-15T14:26:54Z","tags": 
   ["warning","savedobjects-service"],"pid":3176,"message":"Unable to connect to 
    Elasticsearch. Error: Request Timeout after 30000ms"}

this is the last part of the log of elastic cluster:

[2021-01-15T13:58:24,995][INFO ][o.e.c.r.a.AllocationService] [cluster-node-1] Cluster health status changed from [RED] to [GREEN] (reason: [shards started [[index_att][0]]]).
[2021-01-15T14:40:48,800][INFO ][o.e.c.m.MetadataDeleteIndexService] [elastic-cluster-node-1] [.kibana_2/21NiblINQRiAuG3kTEREhA] deleting index
[2021-01-15T14:44:29,509][INFO ][o.e.c.m.MetadataCreateIndexService] [elastic-cluster-node-1] [.kibana_2] creating index, cause [api], templates [], shards [1]/[1], mappings [_doc]
[2021-01-15T14:44:29,525][INFO ][o.e.c.r.a.AllocationService] [cluster-node-1] updating number_of_replicas to [0] for indices [.kibana_2]
[2021-01-15T14:44:29,884][INFO ][o.e.c.r.a.AllocationService] [cluster-node-1] Cluster health status changed from [YELLOW] to [GREEN] (reason: [shards started [[.kibana_2][0]]])

this is elasticsearch.yml

cluster.name: elastic-cluster
node.name: cluster-node-1
path.data: H:\Elastic\data
path.logs: H:\Elastic\logs
network.host: 0.0.0.0
http.port: 9200
discovery.seed_hosts: ["srvenv1", "srvenv2", "srvenv3"]
cluster.initial_master_nodes: ["cluster-node-1", "cluster-node-2", "cluster-node-3"]

Sorry, I've no idea how to recover from a deleted .kibana_1 index. Try asking in the Kibana forum.

1 Like

If you've deleted the index and you don't have a backup, you cannot recover it.

You should be able to restart Kibana and it'll recreate what it needs. If that's not working then create a new topic as David suggested :slight_smile:

1 Like