From 25b82848d610bf87731c08b2551102d15e1ca822 Mon Sep 17 00:00:00 2001
From: Graham McIntire <graham@mcintire.me>
Date: Sat, 3 Jan 2026 16:09:31 -0600
Subject: [PATCH] Run migrations on app start instead of separate job

- Migrations now run in Application.start/2 before starting services
- Ecto's advisory locks prevent concurrent migrations in clustered setup
- Simplified deployment - no separate migration job needed
- Removed migrate-job.yaml manifest
---
 .gitlab-ci.yml              | 13 +-------
 k8s/migrate-job.yaml        | 59 -------------------------------------
 lib/towerops/application.ex |  5 ++++
 3 files changed, 6 insertions(+), 71 deletions(-)
 delete mode 100644 k8s/migrate-job.yaml

diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 94349e8a..69a5031d 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -43,18 +43,7 @@ deploy:
   script:
     - kubectl config get-contexts
     - kubectl config use-context graham/towerops:towerops
-    # Run database migrations before deploying new version
-    - |
-      # Create unique job name using commit SHA (truncated to fit k8s naming limits)
-      JOB_NAME="towerops-migrate-${CI_COMMIT_SHORT_SHA}"
-      # Generate job manifest with correct image tag and unique name
-      sed -e "s|IMAGE_TAG|${CI_COMMIT_SHA}|g" \
-          -e "s|towerops-migrate-TIMESTAMP|${JOB_NAME}|g" \
-          k8s/migrate-job.yaml | kubectl apply -f -
-      # Wait for migration to complete (max 5 minutes)
-      kubectl wait --for=condition=complete --timeout=5m job/${JOB_NAME} -n towerops
-      echo "Database migration completed successfully"
-    # Deploy new version
+    # Deploy new version (migrations run on app start)
     - kubectl set image deployment/towerops towerops=$CI_REGISTRY_IMAGE:$CI_COMMIT_SHA -n towerops
     - kubectl rollout status deployment/towerops -n towerops --timeout=5m
   environment:
diff --git a/k8s/migrate-job.yaml b/k8s/migrate-job.yaml
deleted file mode 100644
index 3d085805..00000000
--- a/k8s/migrate-job.yaml
+++ /dev/null
@@ -1,59 +0,0 @@
----
-apiVersion: batch/v1
-kind: Job
-metadata:
-  name: towerops-migrate-TIMESTAMP
-  namespace: towerops
-  labels:
-    app: towerops
-    component: migration
-spec:
-  # Clean up completed jobs after 1 hour
-  ttlSecondsAfterFinished: 3600
-  backoffLimit: 3
-  template:
-    metadata:
-      labels:
-        app: towerops
-        component: migration
-    spec:
-      restartPolicy: Never
-      imagePullSecrets:
-      - name: gitlab-registry
-      securityContext:
-        runAsNonRoot: true
-        runAsUser: 65534
-        fsGroup: 65534
-        seccompProfile:
-          type: RuntimeDefault
-      containers:
-      - name: migrate
-        image: registry.gitlab.com/graham/towerops:IMAGE_TAG
-        command: ["/app/bin/towerops", "eval", "Towerops.Release.migrate()"]
-        securityContext:
-          allowPrivilegeEscalation: false
-          runAsNonRoot: true
-          runAsUser: 65534
-          capabilities:
-            drop:
-            - ALL
-          seccompProfile:
-            type: RuntimeDefault
-        env:
-        - name: MIX_ENV
-          value: "prod"
-        - name: RELEASE_COOKIE
-          valueFrom:
-            secretKeyRef:
-              name: towerops-secrets
-              key: RELEASE_COOKIE
-        envFrom:
-        - secretRef:
-            name: towerops-db
-        resources:
-          requests:
-            memory: "128Mi"
-            cpu: "100m"
-          limits:
-            memory: "256Mi"
-            cpu: "500m"
diff --git a/lib/towerops/application.ex b/lib/towerops/application.ex
index 78d680d9..a7eca3c7 100644
--- a/lib/towerops/application.ex
+++ b/lib/towerops/application.ex
@@ -7,6 +7,11 @@ defmodule Towerops.Application do
 
   @impl true
   def start(_type, _args) do
+    # Run migrations on startup (Ecto handles locking for concurrent runs)
+    unless Application.get_env(:towerops, Towerops.Repo)[:database] == "towerops_test" do
+      Towerops.Release.migrate()
+    end
+
     topologies = Application.get_env(:libcluster, :topologies, [])
 
     children = [