ECK Cluster in yellow state but cluster is green

[julien.couturier (Couturier)]

Hi,

My Elastic stack is working well but I discovered that the ECK status was "Yellow" when I tried to upgrade to 7.16.2.

kubectl get es
NAME HEALTH NODES VERSION PHASE AGE
tbb-es yellow 3 7.11.2 Ready 34d

But when I check my ES Cluster health, everything is right :

GET _cluster/health

{
  "cluster_name" : "tbb-es",
  "status" : "green",
  "timed_out" : false,
  "number_of_nodes" : 3,
  "number_of_data_nodes" : 3,
  "active_primary_shards" : 312,
  "active_shards" : 761,
  "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
}

GET _cat/allocation?v
shards disk.indices disk.used disk.avail disk.total disk.percent host ip node
253 154.1gb 154.9gb 40.9gb 195.8gb 79 10.2.1.10 10.2.1.10 tbb-es-es-default-2
254 137.7gb 139gb 56.7gb 195.8gb 71 10.2.0.28 10.2.0.28 tbb-es-es-default-0
254 155.2gb 156.5gb 39.2gb 195.8gb 79 10.2.2.8 10.2.2.8 tbb-es-es-default-1

GET _cluster/settings

{
  "persistent" : { },
  "transient" : {
    "cluster" : {
      "routing" : {
        "allocation" : {
          "enable" : "all",
          "exclude" : {
            "_name" : "none_excluded"
          }
        }
      }
    }
  }
}

ECK version is 1.4.0

My update strategy seems good :
Update Strategy:
Change Budget:
Max Surge: 1
Max Unavailable: 1

What can explain this yellow status ?

[michael.morello]

@julien.couturier

What can explain this yellow status ?

Could you check the operator logs to see if there is something preventing the operator to get/update the status ?

Thanks

[julien.couturier (Couturier)]

I already checked the operator logs but I did not find anything significant.
But, on Tuesday I upgraded my kubernetes environment so it was needed to restart all deployments. After this rollup restart, the ECK cluster became Green and started to upgrade itself.

I suppose, a simple rollup restart would have solved my issue too.