[DOP-22054] Move migrations to separated script

Makefile

@@ -130,10 +130,13 @@ dev-worker: db-start broker-start ##@Application Run development broker (without
 
 
 prod-build-server: ##@Application Build docker image for server
-	docker build --progress=plain -t mtsrus/syncmaster-backend:develop -f ./docker/Dockerfile.backend $(ARGS) .
+	docker build --progress=plain -t mtsrus/syncmaster-backend:develop -f ./docker/Dockerfile.backend --target=prod $(ARGS) .
+
+prod-build-scheduler: ##@Application Build docker image for scheduler
+	docker build --progress=plain -t mtsrus/syncmaster-scheduler:develop -f ./docker/Dockerfile.scheduler --target=prod $(ARGS) .
 
 prod-build-worker: ##@Application Build docker image for worker
-	docker build --progress=plain -t mtsrus/syncmaster-worker:develop -f ./docker/Dockerfile.worker $(ARGS) .
+	docker build --progress=plain -t mtsrus/syncmaster-worker:develop -f ./docker/Dockerfile.worker --target=prod $(ARGS) .
 
 prod-build: prod-build-server prod-build-worker ##@Application Build docker images
 

docker-compose.test.yml

@@ -28,6 +28,22 @@ services:
       timeout: 5s
       retries: 3
 
+  migrations:
+    image: mtsrus/syncmaster-worker:${WORKER_IMAGE_TAG:-test}
+    restart: none
+    build:
+      dockerfile: docker/Dockerfile.worker
+      context: .
+      target: test
+    volumes:
+      - ./syncmaster:/app/syncmaster
+      - ./tests:/app/tests
+    entrypoint: [python, -m, syncmaster.db.migrations, upgrade, head]
+    env_file: .env.docker
+    depends_on:
+      db:
+        condition: service_healthy
+
   backend:
     image: mtsrus/syncmaster-backend:${BACKEND_IMAGE_TAG:-test}
     restart: unless-stopped
@@ -41,15 +57,22 @@ services:
     volumes:
       - ./syncmaster:/app/syncmaster
       - ./docs/_static:/app/docs/_static
-      - ./cached_jars:/root/.ivy2
       - ./reports:/app/reports
       - ./tests:/app/tests
       - ./pyproject.toml:/app/pyproject.toml
     depends_on:
       db:
         condition: service_healthy
+      migrations:
+        condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
+    healthcheck:
+      test: [CMD-SHELL, curl -f http://localhost:8000/monitoring/ping]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 5s
     profiles: [backend, all]
 
   scheduler:
@@ -68,6 +91,8 @@ services:
     depends_on:
       db:
         condition: service_healthy
+      migrations:
+        condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
     profiles: [scheduler, all]
@@ -94,6 +119,8 @@ services:
     depends_on:
       db:
         condition: service_healthy
+      migrations:
+        condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
     profiles: [worker, scheduler, s3, oracle, hdfs, hive, clickhouse, mysql, mssql, all]

docker-compose.yml

@@ -28,12 +28,26 @@ services:
       timeout: 5s
       retries: 3
 
+  migrations:
+    image: mtsrus/syncmaster-backend:${TAG:-develop}
+    restart: none
+    build:
+      dockerfile: docker/Dockerfile.backend
+      context: .
+      target: prod
+    entrypoint: [python, -m, syncmaster.db.migrations, upgrade, head]
+    env_file: .env.docker
+    depends_on:
+      db:
+        condition: service_healthy
+
   backend:
     image: mtsrus/syncmaster-backend:${TAG:-develop}
     restart: unless-stopped
     build:
       dockerfile: docker/Dockerfile.backend
       context: .
+      target: prod
     ports:
       - 8000:8000
     environment:
@@ -49,6 +63,8 @@ services:
     depends_on:
       db:
         condition: service_healthy
+      migrations:
+        condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
     healthcheck:
@@ -64,10 +80,13 @@ services:
     build:
       dockerfile: docker/Dockerfile.worker
       context: .
+      target: prod
     env_file: .env.docker
     depends_on:
       db:
         condition: service_healthy
+      migrations:
+        condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
 
@@ -77,10 +96,13 @@ services:
     build:
       dockerfile: docker/Dockerfile.scheduler
       context: .
+      target: prod
     env_file: .env.docker
     depends_on:
       db:
         condition: service_healthy
+      migrations:
+        condition: service_completed_successfully
       rabbitmq:
         condition: service_healthy
 

docker/Dockerfile.backend

@@ -39,7 +39,7 @@ ENV SYNCMASTER__SERVER__OPENAPI__SWAGGER__JS_URL=/static/swagger/swagger-ui-bund
     SYNCMASTER__SERVER__STATIC_FILES__DIRECTORY=/app/syncmaster/backend/static
 
 
-FROM base as test
+FROM base AS test
 
 RUN poetry install --no-root --extras "backend" --with test --without docs,dev
 RUN sed -i 's/python -m/coverage run -m/g' /app/entrypoint.sh

docker/Dockerfile.scheduler

@@ -22,7 +22,8 @@ FROM base AS prod
 
 COPY ./syncmaster/ /app/syncmaster/
 
-FROM base as test
+
+FROM base AS test
 
 RUN poetry install --no-root --all-extras --with test --without docs,dev
-RUN sed -i 's/python -m/coverage run -m/g' /app/entrypoint.sh
+RUN sed -i 's/python -m/coverage run -m/g' /app/entrypoint.sh

docker/Dockerfile.worker

@@ -42,7 +42,7 @@ FROM base as prod
 COPY ./syncmaster/ /app/syncmaster/
 
 
-FROM base as test
+FROM base AS test
 
 ENV SYNCMASTER__WORKER__CREATE_SPARK_SESSION_FUNCTION=tests.spark.get_worker_spark_session
 

docker/entrypoint_backend.sh

@@ -1,8 +1,6 @@
 #!/usr/bin/env bash
 set -e
 
-python -m syncmaster.db.migrations upgrade head
-
 if [[ "x${SYNCMASTER__ENTRYPOINT__SUPERUSERS}" != "x" ]]; then
   superusers=$(echo "${SYNCMASTER__ENTRYPOINT__SUPERUSERS}" | tr "," " ")
   python -m syncmaster.backend.scripts.manage_superusers add ${superusers}

docs/changelog/next_release/163.breaking.rst

@@ -0,0 +1 @@
+Now migrations are executed in a dedicated one-off container, instead of being run as a part of ``backend`` container.

docs/scheduler/start_scheduler.rst

@@ -1,7 +1,6 @@
 Starting the Scheduler
 ======================
 
-
 With docker
 -----------
 
@@ -14,12 +13,12 @@ Options can be set via ``.env`` file or ``environment`` section in ``docker-comp
 .. dropdown:: ``docker-compose.yml``
 
     .. literalinclude:: ../../docker-compose.yml
-       :emphasize-lines: 68-83
+        :emphasize-lines: 90-103
 
 .. dropdown:: ``.env.docker``
 
     .. literalinclude:: ../../.env.docker
-       :emphasize-lines: 11-25
+        :emphasize-lines: 11-25
 
 To start the worker container you need to run the command:
 
@@ -28,7 +27,6 @@ To start the worker container you need to run the command:
     docker compose up scheduler -d --wait --wait-timeout 200
 
 
-
 Without docker
 --------------
 

docs/worker/start_worker.rst

@@ -6,25 +6,24 @@ Starting the Celery Worker
     Before starting the worker you need to create a queue.
     The queue is created by sending a post request to ``/queues`` endpoint (See Swagger doc for details).
 
-
 With docker
 -----------
 
 Installation process
 ~~~~~~~~~~~~~~~~~~~~
 
-Docker will download worker image of syncmaster worker & broker, and run them.
+Docker will download worker image, and then start worker container with dependencies.
 Options can be set via ``.env`` file or ``environment`` section in ``docker-compose.yml``
 
 .. dropdown:: ``docker-compose.yml``
 
     .. literalinclude:: ../../docker-compose.yml
-       :emphasize-lines: 55-66
+        :emphasize-lines: 75-88
 
 .. dropdown:: ``.env.docker``
 
     .. literalinclude:: ../../.env.docker
-       :emphasize-lines: 11-22
+        :emphasize-lines: 11-22
 
 To start the worker container you need to run the command:
 
@@ -33,21 +32,20 @@ To start the worker container you need to run the command:
     docker compose up worker -d --wait --wait-timeout 200
 
 
-
 Without docker
 --------------
 
 To start the worker you need to run the command
 
 .. code-block:: bash
 
-    python -m celery -A syncmaster.worker.celery worker
+    python -m celery -A syncmaster.worker.celery worker --max-tasks-per-child=1
 
 You can specify options like concurrency and queues by adding additional flags:
 
 .. code-block:: bash
 
-    celery -A -A syncmaster.worker.celery  worker --concurrency=4 --max-tasks-per-child=1 --loglevel=info
+    python -m celery -A syncmaster.worker.celery worker --concurrency=4 --max-tasks-per-child=1 --loglevel=info
 
 
 Refer to the `Celery <https://docs.celeryq.dev/en/stable/>`_ documentation for more advanced start options.

pyproject.toml

@@ -71,7 +71,6 @@ apscheduler = { version = "^3.10.4", optional = true }
 starlette-exporter = {version = "^0.23.0", optional = true}
 itsdangerous = {version = "*", optional = true}
 python-keycloak = {version = "^4.7.0", optional = true}
-devtools = {version = "*", optional = true}
 
 [tool.poetry.extras]
 backend = [
@@ -91,7 +90,6 @@ backend = [
     "coloredlogs",
     "python-json-logger",
     "asyncpg",
-    "devtools",
     "itsdangerous",
     "python-keycloak",
     "pyyaml",
@@ -110,7 +108,6 @@ worker = [
     "uuid6",
     "coloredlogs",
     "python-json-logger",
-    "python-keycloak",
 ]
 
 scheduler = [

syncmaster/backend/providers/auth/dummy_provider.py

@@ -2,10 +2,10 @@
 # SPDX-License-Identifier: Apache-2.0
 
 import logging
+from pprint import pformat
 from time import time
 from typing import Annotated, Any
 
-from devtools import pformat
 from fastapi import Depends, FastAPI
 
 from syncmaster.backend.dependencies import Stub