recovery(metis): seed per-node vault password slots

services/keycloak/kustomization.yaml

@@ -26,6 +26,7 @@ resources:
   - oneoffs/soteria-oidc-secret-ensure-job.yaml
   - oneoffs/quality-oidc-secret-ensure-job.yaml
   - oneoffs/metis-ssh-keys-secret-ensure-job.yaml
+  - oneoffs/metis-node-passwords-secret-ensure-job.yaml
   - oneoffs/harbor-oidc-secret-ensure-job.yaml
   - oneoffs/vault-oidc-secret-ensure-job.yaml
   - oneoffs/actual-oidc-secret-ensure-job.yaml

services/keycloak/oneoffs/metis-node-passwords-secret-ensure-job.yaml

@@ -0,0 +1,114 @@
+# services/keycloak/oneoffs/metis-node-passwords-secret-ensure-job.yaml
+# One-off job for sso/metis-node-passwords-secret-ensure-1.
+# Purpose: ensure per-node Metis recovery password placeholders exist in Vault.
+# Existing values are preserved; only missing fields are initialized.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: metis-node-passwords-secret-ensure-1
+  namespace: sso
+spec:
+  backoffLimit: 0
+  ttlSecondsAfterFinished: 3600
+  template:
+    spec:
+      serviceAccountName: mas-secrets-ensure
+      restartPolicy: Never
+      affinity:
+        nodeAffinity:
+          requiredDuringSchedulingIgnoredDuringExecution:
+            nodeSelectorTerms:
+              - matchExpressions:
+                  - key: node-role.kubernetes.io/worker
+                    operator: Exists
+          preferredDuringSchedulingIgnoredDuringExecution:
+            - weight: 100
+              preference:
+                matchExpressions:
+                  - key: kubernetes.io/arch
+                    operator: In
+                    values: ["arm64"]
+      containers:
+        - name: apply
+          image: registry.bstein.dev/bstein/kubectl:1.35.0
+          command: ["/bin/sh", "-c"]
+          args:
+            - |
+              set -eu
+
+              vault_addr="${VAULT_ADDR:-http://vault.vault.svc.cluster.local:8200}"
+              vault_role="${VAULT_ROLE:-sso-secrets}"
+
+              jwt="$(cat /var/run/secrets/kubernetes.io/serviceaccount/token)"
+              login_payload="$(jq -nc --arg jwt "${jwt}" --arg role "${vault_role}" '{jwt:$jwt, role:$role}')"
+              vault_token="$(curl -sS --request POST --data "${login_payload}"                 "${vault_addr}/v1/auth/kubernetes/login" | jq -r '.auth.client_token')"
+              if [ -z "${vault_token}" ] || [ "${vault_token}" = "null" ]; then
+                echo "vault login failed" >&2
+                exit 1
+              fi
+
+              nodes="
+              titan-04
+              titan-05
+              titan-06
+              titan-07
+              titan-08
+              titan-09
+              titan-0a
+              titan-0b
+              titan-0c
+              titan-10
+              titan-11
+              titan-12
+              titan-13
+              titan-14
+              titan-15
+              titan-16
+              titan-17
+              titan-18
+              titan-19
+              titan-20
+              titan-21
+              titan-22
+              titan-23
+              titan-24
+              "
+
+              ensured=0
+              for node in ${nodes}; do
+                secret_path="secret/data/nodes/${node}"
+                read_status="$(curl -sS -o /tmp/node-read.json -w "%{http_code}"                   -H "X-Vault-Token: ${vault_token}"                   "${vault_addr}/v1/${secret_path}" || true)"
+                if [ "${read_status}" = "200" ]; then
+                  ssh_password="$(jq -r '.data.data.ssh_password // empty' /tmp/node-read.json)"
+                  ssh_password_hash="$(jq -r '.data.data.ssh_password_hash // empty' /tmp/node-read.json)"
+                  atlas_password="$(jq -r '.data.data.atlas_password // empty' /tmp/node-read.json)"
+                  atlas_password_hash="$(jq -r '.data.data.atlas_password_hash // empty' /tmp/node-read.json)"
+                  root_password="$(jq -r '.data.data.root_password // empty' /tmp/node-read.json)"
+                  root_password_hash="$(jq -r '.data.data.root_password_hash // empty' /tmp/node-read.json)"
+                elif [ "${read_status}" = "404" ]; then
+                  ssh_password=""
+                  ssh_password_hash=""
+                  atlas_password=""
+                  atlas_password_hash=""
+                  root_password=""
+                  root_password_hash=""
+                else
+                  echo "Vault read failed for ${node} (status ${read_status})" >&2
+                  cat /tmp/node-read.json >&2 || true
+                  exit 1
+                fi
+
+                payload="$(jq -nc                   --arg hostname "${node}"                   --arg ssh_password "${ssh_password}"                   --arg ssh_password_hash "${ssh_password_hash}"                   --arg atlas_password "${atlas_password}"                   --arg atlas_password_hash "${atlas_password_hash}"                   --arg root_password "${root_password}"                   --arg root_password_hash "${root_password_hash}"                   '{data:{hostname:$hostname,ssh_password:$ssh_password,ssh_password_hash:$ssh_password_hash,atlas_password:$atlas_password,atlas_password_hash:$atlas_password_hash,root_password:$root_password,root_password_hash:$root_password_hash}}')"
+
+                write_status="$(curl -sS -o /tmp/node-write.json -w "%{http_code}" -X POST                   -H "X-Vault-Token: ${vault_token}"                   -H 'Content-Type: application/json'                   -d "${payload}"                   "${vault_addr}/v1/${secret_path}")"
+                if [ "${write_status}" != "200" ] && [ "${write_status}" != "204" ]; then
+                  echo "Vault write failed for ${node} (status ${write_status})" >&2
+                  cat /tmp/node-write.json >&2 || true
+                  exit 1
+                fi
+
+                ensured=$((ensured + 1))
+                echo "Ensured node secret placeholder for ${node}"
+              done
+
+              echo "Ensured ${ensured} Metis node password placeholders in Vault"

services/vault/scripts/vault_k8s_auth_configure.sh

@@ -87,6 +87,7 @@ write_policy_and_role() {
   service_accounts="$3"
   read_paths="$4"
   write_paths="$5"
+  extra_rules="${6:-}"
 
   policy_body=""
   for path in ${read_paths}; do
@@ -109,6 +110,11 @@ path \"kv/metadata/atlas/${path}\" {
 }
 "
   done
+  if [ -n "${extra_rules}" ]; then
+    policy_body="${policy_body}
+${extra_rules}
+"
+  fi
 
   log "writing policy ${role}"
   printf '%s\n' "${policy_body}" | vault_cmd policy write "${role}" -
@@ -231,7 +237,15 @@ write_policy_and_role "crypto" "crypto" "crypto-vault-sync" \
 write_policy_and_role "health" "health" "health-vault-sync" \
   "health/*" ""
 write_policy_and_role "maintenance" "maintenance" "ariadne,maintenance-vault-sync,metis" \
-  "maintenance/ariadne-db maintenance/metis-oidc maintenance/soteria-oidc maintenance/metis-ssh-keys maintenance/metis-runtime portal/atlas-portal-db portal/bstein-dev-home-keycloak-admin mailu/mailu-db-secret mailu/mailu-initial-account-secret nextcloud/nextcloud-db nextcloud/nextcloud-admin health/wger-admin finance/firefly-secrets comms/mas-admin-client-runtime comms/atlasbot-credentials-runtime comms/synapse-db comms/synapse-admin vault/vault-oidc-config shared/harbor-pull harbor/harbor-core" ""
+  "maintenance/ariadne-db maintenance/metis-oidc maintenance/soteria-oidc maintenance/metis-ssh-keys maintenance/metis-runtime portal/atlas-portal-db portal/bstein-dev-home-keycloak-admin mailu/mailu-db-secret mailu/mailu-initial-account-secret nextcloud/nextcloud-db nextcloud/nextcloud-admin health/wger-admin finance/firefly-secrets comms/mas-admin-client-runtime comms/atlasbot-credentials-runtime comms/synapse-db comms/synapse-admin vault/vault-oidc-config shared/harbor-pull harbor/harbor-core" "" \
+  '
+path "secret/data/nodes/*" {
+  capabilities = ["read"]
+}
+path "secret/metadata/nodes/*" {
+  capabilities = ["list"]
+}
+'
 write_policy_and_role "maintenance-metis-token-sync" "maintenance" "metis-token-sync" \
   "" \
   "maintenance/metis-runtime"
@@ -249,7 +263,15 @@ write_policy_and_role "vault" "vault" "vault" \
 
 write_policy_and_role "sso-secrets" "sso" "mas-secrets-ensure" \
   "shared/keycloak-admin maintenance/metis-ssh-keys" \
-  "harbor/harbor-oidc vault/vault-oidc-config comms/synapse-oidc logging/oauth2-proxy-logs-oidc finance/actual-oidc maintenance/metis-oidc maintenance/soteria-oidc maintenance/metis-ssh-keys"
+  "harbor/harbor-oidc vault/vault-oidc-config comms/synapse-oidc logging/oauth2-proxy-logs-oidc finance/actual-oidc maintenance/metis-oidc maintenance/soteria-oidc maintenance/metis-ssh-keys" \
+  '
+path "secret/data/nodes/*" {
+  capabilities = ["create", "update", "read"]
+}
+path "secret/metadata/nodes/*" {
+  capabilities = ["list"]
+}
+'
 write_policy_and_role "crypto-secrets" "crypto" "crypto-secrets-ensure" \
   "" \
   "crypto/wallet-monero-temp-rpc-auth"