AI search engine and MCP server for bots and agents. One Rust binary indexes code, docs, logs, JSONL/CSV and PDFs for BM25 full-text, vector and hybrid search, code search without grep, agent memory and RAG. Elasticsearch-compatible API; Corpus Hub serves hundreds of reference memory corpora.
Rust agent-memoryai-agentsbm25claude-codeclicode-searchcursordatabaseelasticsearchfull-text-searchhybrid-searchjsonllocal-firstmcpmcp-servermodel-context-protocolragrustsearch-enginevector-search
fix(helm): default to secure, and say what insecure mode turns off
xerj-org3e2a0a14d72c
diff --git a/deploy/helm/xerj/templates/NOTES.txt b/deploy/helm/xerj/templates/NOTES.txtnew file mode 100644index 000000000..6eb69b013--- /dev/null+++ b/deploy/helm/xerj/templates/NOTES.txt@@ -0,0 +1,42 @@+XERJ {{ .Chart.AppVersion }} installed as {{ .Release.Name }} in namespace {{ .Release.Namespace }}.++{{- if .Values.insecure }}++ WARNING: insecure mode is ON.++ TLS and authentication are both disabled. Every listener below is open to+ anything that can reach this pod:++ ES-compat {{ .Values.service.esCompatPort }}+ native REST {{ .Values.service.nativePort }}+ gRPC {{ .Values.service.grpcPort }}++ This is fine for a local dev cluster and is not fine for anything holding+ real data. Turn it off with:++ helm upgrade {{ .Release.Name }} ... --set insecure=false++{{- else }}++ Auth is on. XERJ generates an admin key on first start and writes it to+ {{ .Values.persistence.mountPath | default "/var/lib/xerj" }}/admin.key with mode 0600.++ Read it with:++ kubectl exec -n {{ .Release.Namespace }} {{ .Release.Name }}-0 -- cat {{ .Values.persistence.mountPath | default "/var/lib/xerj" }}/admin.key++ Then call the API with:++ curl -H "Authorization: ApiKey $KEY" http://{{ .Release.Name }}:{{ .Values.service.esCompatPort }}/_cluster/health++{{- if not .Values.tls.enabled }}++ Note: TLS is off, so that key crosses the network in cleartext. Set+ tls.enabled=true and supply tls.secretName before exposing this outside+ the cluster.+{{- end }}+{{- end }}++Check it is up:++ kubectl get pods -n {{ .Release.Namespace }} -l app.kubernetes.io/instance={{ .Release.Name }}diff --git a/deploy/helm/xerj/values.yaml b/deploy/helm/xerj/values.yamlindex 5fde2f641..c9211fb0b 100644--- a/deploy/helm/xerj/values.yaml+++ b/deploy/helm/xerj/values.yaml@@ -50,9 +50,16 @@ probes:timeoutSeconds: 2failureThreshold: 3-# Insecure mode disables TLS + auth (for dev/POC clusters). Production-# users should set this to false and supply tls.cert/key + auth.token.-insecure: true+# Insecure mode disables TLS AND auth. Off by default: an install with no+# values.yaml of your own is authenticated, matching the engine's own default+# (auth.enabled = true, admin key auto-generated 0600 on first start).+#+# Set to true ONLY for a local dev or throwaway POC cluster. Doing so leaves+# the ES-compat, native REST and gRPC listeners open to anything that can reach+# the pod. Insecure defaults are the single most cited cause of the NoSQL data+# exposures in user-feedback/09-security/insecure-defaults.md, and a chart+# default is the configuration most users will ever run.+insecure: falsetls:enabled: false