ElasticSearch with Kubernetes

Code snippets and notes on running Apache Spark with Kubernetes.

Installation

Install Elastic Cloud on Kubbernetes Operator

kubectl create -f crds.yaml
kubectl create -f operator.yaml

# confirm operator installation
kubectl -n elastic-system get pod

# Monitor the operator logs:
kubectl -n elastic-system logs -f statefulset.apps/elastic-operator

Create a Elastic cluster:

kubectl apply -f reddot.yaml

Browser to Kibana UI at kibana.purestorage.int

Get elastic user password.

kubectl -n bds get secret reddot-es-elastic-user -o=jsonpath='{.data.elastic}' | base64 --decode; echo

Visualize k8s logs using ES

Create index partern: k8s-logs and k8s-systemd. Go to Kibana - Stack Management - Index Patterns

Import k8s-logs dashboard using export.ndjson. Go to Kibana - Stack Management - Saved objects.

Snapshot to S3

Reference:

Install S3 snapshot repository

Create a secret for S3 access keys:

kubectl create secret generic es-s3-keys -n elastic --from-literal=access-key='xxxx' --from-literal=secret-key='vvvv'

Install S3 repository plugin:

kubectl apply -f reddot.yaml

Go to Kibana Stack Management UI to register the repository.

Create S3 repository

Create a FB S3 repository:

PUT _snapshot/reddot-s3-repo?pretty
  {
  "type": "s3",
  "settings": {
    "bucket": "deephub",
    "base_path": "elastic/snapshots",
    "endpoint": "192.168.170.11",
    "protocol": "http",
    "max_restore_bytes_per_sec": "1gb",
    "max_snapshot_bytes_per_sec": "200mb"
  }
}

Check repositories on Repositories UI.

Create snapshots

Create a snapshot policy on Stack Management UI.

Demo

Demo flow:

k8s cluster walk through, PSO storage class
k8s logging with Elastic & FB: search 'error', dashboard
Scale Elastic: data node count, pure-file, FB UI
Snapshot to S3: repo setting, s3 ls

s3 ls s3://deephub/elastic/snapshots/

uprush/kube-elastic