LSDmop Helm Chart

Quick and Dirty Install

helm repo add lsdopen https://lsdopen.github.io/lsdmop
helm repo update

Install on Generic Kubernetes Cluster

Basic installation

Only run kubectl scale sts lsdmop-logstash --replicas=1 once elastic-post-setup has finished running

helm show values lsdopen/lsdmop > values.yaml
helm install lsdmop lsdopen/lsdmop -n lsdmop --create-namespace --values values.yaml ; helm upgrade lsdmop lsdopen/lsdmop -n lsdmop --values values.yaml

kubectl scale sts lsdmop-logstash --replicas=1

Install on GKE with Admin Password and Elastic disabled

helm install lsdmop -n lsdmop --create-namespace \
--set lsdmop.clusterType=gke \
--set lsdmop.elastic.enabled=false \
--set grafana.adminPassword=neiliscool

Install on Rancher with Admin Password and SMTP details and Elastic disabled

helm install lsdmop -n lsdmop --create-namespace \
--set lsdmop.clusterType=rancher \
--set lsdmop.elastic.enabled=false \
--set grafana.adminPassword=neiliscool \
--set lsdmop.smtp.host=smtp.lsdopen.io \
--set lsdmop.smtp.port=25 \
--set lsdmop.smtp.fromAddress="lsdmop+noreply+api.mycluster.lsdopen.io@lsdopen.io" \
--set lsdmop.smtp.fromName="LSDobserve - LSD - api.mycluster.lsdopen.io"

Restricted Network Installation

Simple installation

Only run kubectl scale sts lsdmop-logstash --replicas=1 once elastic-post-setup has finished running

helm dependency update .

kubectl create ns lsdmop
helm install lsdmop . -n lsdmop --create-namespace --values values.yaml ; helm upgrade lsdmop . -n lsdmop --create-namespace --values values.yaml

kubectl scale sts lsdmop-logstash --replicas=1

Test smtp details

Run this to trigger an alert to alertmanager manually: helm test lsdmop

Adding New Dashboards to Grafana

1. Place new dashboards in ./dashboards folder
2. Modify templates/grafana.dashbaords.yaml with your new dashboard details
3. Add dashboard to templates/grafana.configmap.yaml
4. Add the dashboard to the values.yaml file

Todo

• harbor dashboards
• alertmanager dashbaord
• Dashboards in Elastic or Elastic like dashboards in Grafana
• Openshift - patch namespace to allow it to run on any nodes "oc patch namespace lsdmop -p '{"metadata":{"annotations":{"openshift.io/node-selector":""}}}'"

Notes

If you get the following error on Filebeat

2021-03-03T12:05:09.464Z	INFO	instance/beat.go:645	Home path: [/usr/share/filebeat] Config path: [/usr/share/filebeat] Data path: [/usr/share/filebeat/data] Logs path: [/usr/share/filebeat/logs]
2021-03-03T12:05:09.464Z	INFO	instance/beat.go:373	filebeat stopped.
2021-03-03T12:05:09.464Z	ERROR	instance/beat.go:956	Exiting: Failed to create Beat meta file: open /usr/share/filebeat/data/meta.json.new: permission denied
Exiting: Failed to create Beat meta file: open /usr/share/filebeat/data/meta.json.new: permission denied

Then you have an SELinux issue, and can resolve it with the following command, that needs to be run on all servers

chcon -t container_file_t /var/lib/lsdmop/lsdmop-filebeat/filebeat-data

Delete the elastic-post-setup job before upgrading - otherwise the correcnt tempaltes and ILM policies will not be applied.

Manual removal

k delete ns lsdmop ; k delete clusterrole lsdmop-kube-state-metrics elastic-beat-autodiscover elastic-operator elastic-operator-edit elastic-operator-view lsdmop-grafana-clusterrole lsdmop-kube-state-metrics lsdmop-prometheus-alertmanager lsdmop-prometheus-server ; k delete psp lsdmop-grafana lsdmop-grafana-test lsdmop-prometheus-blackbox-exporter-psp ; kubectl delete ClusterRoleBinding lsdmop-grafana-clusterrolebinding lsdmop-kube-state-metrics lsdmop-prometheus-alertmanager lsdmop-prometheus-server  ; k delete ClusterRoleBinding elastic-operator elastic-beat-autodiscover-binding