updated dockerfile to support multi arch during build and publish and…

.github/workflows/verify.yml

@@ -0,0 +1,91 @@
+name: Verify (multi-arch build + smoke)
+
+on:
+  pull_request:
+  workflow_dispatch:
+  push:
+    branches:
+      - main
+      - master
+
+jobs:
+  build:
+    name: Build multi-arch (no push)
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: appwrite/ollama
+          tags: |
+            type=sha,format=short
+
+      - name: Build (amd64 + arm64)
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: ./dockerfile
+          platforms: linux/amd64,linux/arm64
+          push: false
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          build-args: |
+            VERSION=${{ steps.meta.outputs.version }}
+            MODELS=
+            OLLAMA_KEEP_ALIVE=1m
+
+  smoke:
+    name: Smoke test (docker compose + embeddings)
+    runs-on: ubuntu-latest
+    env:
+      _APP_EMBEDDING_MODELS: embeddinggemma
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+      - name: docker compose up (build + start)
+        run: |
+          docker compose -f docker-compose.yml up --build -d
+
+      - name: Wait for server
+        run: |
+          for i in $(seq 1 30); do
+            if curl -fsS http://127.0.0.1:11434/api/version >/dev/null; then
+              exit 0
+            fi
+            sleep 1
+          done
+          echo "Ollama did not become ready in time"
+          docker logs ollama-verify || true
+          exit 1
+
+      - name: Show version
+        run: curl -fsS http://127.0.0.1:11434/api/version
+
+      - name: Test embeddinggemma embeddings
+        run: |
+          curl -fsS http://127.0.0.1:11434/api/embeddings \
+            -H 'Content-Type: application/json' \
+            -d '{
+              "model": "embeddinggemma",
+              "prompt": "Hello from embeddinggemma"
+            }'
+
+      - name: Failure logs
+        if: failure()
+        run: |
+          echo "=== Docker Compose Logs ==="
+          docker compose -f docker-compose.yml logs || true
+
+      - name: Cleanup
+        if: always()
+        run: docker compose -f docker-compose.yml down -v || true

dockerfile

@@ -9,16 +9,43 @@ ARG OLLAMA_KEEP_ALIVE
 ENV OLLAMA_KEEP_ALIVE=${OLLAMA_KEEP_ALIVE:-24h}
 
 # Pre-pull models at build time for Docker layer caching
-RUN ollama serve & \
-    sleep 5 && \
-    for m in $MODELS; do \
-        echo "Pulling model $m..."; \
-        ollama pull "$m" || exit 1; \
-    done && \
-    pkill ollama
+ARG TARGETARCH
+RUN if [ -z "${MODELS:-}" ]; then \
+        echo "MODELS is empty; skipping build-time pre-pull."; \
+    else \
+        case "${TARGETARCH:-}" in \
+            amd64) PORT=11434 ;; \
+            arm64) PORT=11435 ;; \
+            *) PORT=11436 ;; \
+        esac; \
+        export OLLAMA_HOST="http://127.0.0.1:${PORT}"; \
+        ollama serve >/tmp/ollama-serve.log 2>&1 & pid="$!"; \
+        ready=0; \
+        for i in 1 2 3 4 5 6 7 8 9 10; do \
+            if ollama list >/dev/null 2>&1; then \
+                ready=1; \
+                break; \
+            fi; \
+            sleep 1; \
+        done; \
+        if [ "$ready" -ne 1 ]; then \
+            echo "ERROR: ollama did not become ready during build-time pre-pull" >&2; \
+            echo "--- /tmp/ollama-serve.log ---" >&2; \
+            cat /tmp/ollama-serve.log >&2 || true; \
+            kill "$pid" || true; \
+            wait "$pid" || true; \
+            exit 1; \
+        fi; \
+        for m in $MODELS; do \
+            echo "Pulling model $m..."; \
+            ollama pull "$m" || exit 1; \
+        done; \
+        kill "$pid"; \
+        wait "$pid" || true; \
+    fi
 
 # Expose Ollama default port
 EXPOSE 11434
 
 # On container start, quickly ensure models exist (no re-download unless missing)
-ENTRYPOINT ["/bin/bash", "-c", "read -r -a models <<< \"$MODELS\"; for m in \"${models[@]}\"; do ollama list | grep -qw \"$m\" || ollama pull \"$m\" || exit 1; done && exec ollama serve"]
+ENTRYPOINT ["/bin/bash", "-lc", "set -euo pipefail; if [[ -n \"${MODELS:-}\" ]]; then read -r -a models <<< \"${MODELS}\"; ollama serve >/tmp/ollama-entrypoint.log 2>&1 & pid=$!; ready=0; for i in {1..20}; do if ollama list >/dev/null 2>&1; then ready=1; break; fi; sleep 1; done; if [[ \"$ready\" -ne 1 ]]; then echo 'ERROR: ollama did not become ready during entrypoint pre-pull' >&2; echo '--- /tmp/ollama-entrypoint.log ---' >&2; cat /tmp/ollama-entrypoint.log >&2 || true; kill \"$pid\" || true; wait \"$pid\" || true; exit 1; fi; for m in \"${models[@]}\"; do ollama list | grep -qw \"$m\" || ollama pull \"$m\"; done; kill \"$pid\"; wait \"$pid\" || true; fi; exec ollama serve"]