@onlineapps/conn-orch-validator 12.1.1 → 12.2.0

package/CHANGELOG.md

@@ -4,6 +4,16 @@ All notable changes to this package. Follows [Keep a Changelog](https://keepacha
 
 ## [Unreleased]
 
+### Changed — šablona biz CI: obraz se staví jen na `main`, produkce nasazuje digest z registru (d.671, d.671b; conf image-promotion 001)
+
+Job `build` v `templates/business-service/.gitlab-ci.yml` běží jen na `main` a před stavbou se ptá
+registru, zda tag `:$CI_COMMIT_SHA` už existuje (existuje → nestaví a vypíše digest); tag se nikdy
+nepřepíše. `deploy-production` nebere digest z dotenv artefaktu (`dependencies: []`), ale týmž
+dotazem do registru (`.oa-registry-digest`); když obraz pro commit nenajde, zastaví nasazení
+jmenovanou hláškou a nikdy nestaví náhradu. Fixtury manifestu nesou tentýž blok `oa-ci v1`;
+tvar i tělo dotazu měří `tests/unit/templateImagePromotion.test.js`
+a `tests/unit/templateImagePromotionShell.integration.test.js`.
+
 ## [12.1.1] — 2026-09-25
 
 ### Fixed — `verify-deploy-uniform.sh`: rada při checkoutu mimo `api_biz/` vede na cestu, kterou šablona skutečně deklaruje (d.899)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@onlineapps/conn-orch-validator",
-  "version": "12.1.1",
+  "version": "12.2.0",
   "description": "Validation orchestrator for OA Drive microservices - coordinates validation across all layers (base, infra, orch, business)",
   "oa": {
     "category": "orchestration"

package/templates/business-service/.gitlab-ci.yml

@@ -1,15 +1,17 @@
 # --- oa-ci v1
 # Written by @onlineapps/conn-orch-validator (row G-CI of
 # manifests/biz-service.manifest.json). The PLATFORM half of this pipeline: which
-# pipelines run at all, how a service is built, how its image is identified (R5),
-# how the digest reaches the deploy (R1/R2), the uniform gate that runs on every
-# pipeline (job validate-uniform, confirmation biz-service-manifest 010) and
-# again before the SSH step as the binding instance (008), and the post-deploy
-# gate that runs after it (confirmation deploy-gate-targets 003). The
-# installation contract is checked by that uniform gate, in the manifest rows
-# G-SETUP, D-DB-PACKAGE and D-DB-HEADERS, and no longer by a job of its own
-# (d.470). npx oa-sync-template .gitlab-ci.yml --target . rewrites everything
-# between these two markers and nothing outside them.
+# pipelines run at all, how a service is built - ONCE, on main, never again on
+# production (confirmation image-promotion 001) - how its image is identified
+# (R5), how that one image's digest reaches the deploy (R1/R2), the uniform gate
+# that runs on every pipeline (job validate-uniform, confirmation
+# biz-service-manifest 010) and again before the SSH step as the binding instance
+# (008), and the post-deploy gate that runs after it (confirmation
+# deploy-gate-targets 003). The installation contract is checked by that uniform
+# gate, in the manifest rows G-SETUP, D-DB-PACKAGE and D-DB-HEADERS, and no
+# longer by a job of its own (d.470). npx oa-sync-template .gitlab-ci.yml
+# --target . rewrites everything between these two markers and nothing outside
+# them.
 #
 # Outside them is this repository's own: its test job - which database, which
 # ci:gate:* steps and which artefacts its integration needs is a fact about the
@@ -87,6 +89,126 @@ variables:
     # runner this job carries onto the box comes from that same clone.
     API_UNIFORM_REF_DEPLOY: production
 
+# The ONE question this pipeline asks the container registry (confirmation
+# image-promotion 001): is $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA there, and under
+# which digest? Two jobs need the answer for two different reasons - build, so
+# that a tag already written is never overwritten; deploy-production, because a
+# dotenv artefact does not cross pipelines and production deploys the image main
+# built - and both read it from the registry, the only place both pipelines can
+# see. One question, one implementation: a second copy of this query would be a
+# second rail over one fact (.claude/rules/change-discipline.md § One rail per
+# concern), and the two would answer differently the day one of them was fixed.
+#
+# A hidden key is a declaration and never a job - GitLab runs nothing whose name
+# starts with a dot - so both jobs pull the body in with !reference.
+#
+#   In:  OA_DIGEST_MODE - who is asking, because a missing image means something
+#        different to each of them. "optional": the build's first question, and
+#        an absent tag is the ordinary answer. "build": the build's binding read
+#        after it has built or reused, where an absent tag is a defect. "deploy":
+#        production, where an absent tag is a deploy that must not happen.
+#   Out: OA_IMAGE_PRESENT   - "yes" or "no"
+#        BIZ_IMAGE_DIGEST   - sha256:<64 hex> when present, empty when not.
+#
+# A boolean would not do: the two callers that REQUIRE an image need two
+# different sentences, and one message covering both would name neither the
+# situation nor the fix (architecture-principles.md §5).
+#
+# curl and jq are a declared precondition, not an assumption: the body checks for
+# them before it asks anything (automation-gates.md §2). The deploy job installs
+# them for the post-deploy gate already; the build job's image is alpine and
+# carries neither, so it adds them in a before_script of its own.
+.oa-registry-digest:
+  script:
+    - |
+      # oa-registry-digest
+      for oa_tool in curl jq; do
+        if ! command -v "$oa_tool" > /dev/null 2>&1; then
+          echo "[registry] FATAL: $oa_tool is not on PATH, and the image of this commit can only be found by asking the registry. Expected: curl and jq in this job's image. Fix: add 'apk add --no-cache curl jq' to this job's before_script." >&2
+          exit 1
+        fi
+      done
+      case "${OA_DIGEST_MODE:-}" in
+        optional|build|deploy) ;;
+        *)
+          echo "[registry] FATAL: OA_DIGEST_MODE is '${OA_DIGEST_MODE:-}', and this query has to know what a commit with no image means to the job asking. Expected: 'optional' (the build's first question), 'build' (its binding read after building) or 'deploy' (production). Fix: declare OA_DIGEST_MODE in this job's variables." >&2
+          exit 1 ;;
+      esac
+      # A pull token for THIS project, from this job's own CI_JOB_TOKEN. Both
+      # calls take their credentials AND their URL on stdin (--config -): an
+      # argument stands in the runner's process table for the length of the call,
+      # which is the same reason the registry login below takes its password
+      # there rather than on a command line.
+      oa_exit=0
+      oa_token_json=$(curl --silent --show-error --fail --config - <<CFG
+      user = "gitlab-ci-token:$CI_JOB_TOKEN"
+      url = "$CI_SERVER_URL/jwt/auth?service=container_registry&scope=repository:$CI_PROJECT_PATH:pull"
+      CFG
+      ) || oa_exit=$?
+      if [ "$oa_exit" != "0" ]; then
+        echo "[registry] FATAL: no pull token for $CI_PROJECT_PATH - $CI_SERVER_URL/jwt/auth exited $oa_exit. Expected: a JSON body carrying .token for scope repository:$CI_PROJECT_PATH:pull. Fix: re-run once GitLab answers; this job never continues without a token." >&2
+        exit 1
+      fi
+      oa_token=$(printf '%s' "$oa_token_json" | jq -r '.token // empty')
+      if [ -z "$oa_token" ]; then
+        echo "[registry] FATAL: the token endpoint answered without a .token for $CI_PROJECT_PATH. Expected: {\"token\":\"…\"} for scope repository:$CI_PROJECT_PATH:pull. Fix: check that this project's CI job token may pull its own registry." >&2
+        exit 1
+      fi
+      # HEAD, because the digest is a HEADER: the registry answers
+      # Docker-Content-Digest for a tag without sending the manifest at all. All
+      # four media types are offered, so a single-platform image and an index
+      # both answer with their own digest instead of 406.
+      oa_headers=$(mktemp)
+      oa_exit=0
+      oa_status=$(curl --silent --show-error --head --output /dev/null \
+        --dump-header "$oa_headers" --write-out '%{http_code}' --config - <<CFG
+      header = "Authorization: Bearer $oa_token"
+      header = "Accept: application/vnd.docker.distribution.manifest.v2+json"
+      header = "Accept: application/vnd.oci.image.manifest.v1+json"
+      header = "Accept: application/vnd.docker.distribution.manifest.list.v2+json"
+      header = "Accept: application/vnd.oci.image.index.v1+json"
+      url = "https://$CI_REGISTRY/v2/$CI_PROJECT_PATH/manifests/$CI_COMMIT_SHA"
+      CFG
+      ) || oa_exit=$?
+      if [ "$oa_exit" != "0" ]; then
+        rm -f "$oa_headers"
+        echo "[registry] FATAL: the registry did not answer - HEAD https://$CI_REGISTRY/v2/$CI_PROJECT_PATH/manifests/$CI_COMMIT_SHA, curl exited $oa_exit. Expected: HTTP 200 (the tag is there) or 404 (it is not). Fix: re-run this job once $CI_REGISTRY is reachable - a registry that could not answer is not a registry that said no." >&2
+        exit 1
+      fi
+      # Header names are case-insensitive and the value is lowercase hex either
+      # way, so the whole file is lowercased rather than the match loosened.
+      oa_digest=$(tr '[:upper:]' '[:lower:]' < "$oa_headers" | sed -n 's/^docker-content-digest:[[:space:]]*//p' | tr -d '\r' | head -n 1)
+      rm -f "$oa_headers"
+      case "$oa_status" in
+        200)
+          BIZ_IMAGE_DIGEST="$oa_digest"
+          if ! printf '%s' "${BIZ_IMAGE_DIGEST:-}" | grep -qE '^sha256:[0-9a-f]{64}$'; then
+            echo "[registry] FATAL: $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA is in the registry and its Docker-Content-Digest reads '$BIZ_IMAGE_DIGEST' - expected sha256:<64 hex>, the immutable target R1 pins the production compose to. Fix: re-run this job; if it repeats, the registry is answering a shape this platform does not deploy." >&2
+            exit 1
+          fi
+          OA_IMAGE_PRESENT=yes
+          echo "[registry] $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA -> $BIZ_IMAGE_DIGEST" ;;
+        404)
+          OA_IMAGE_PRESENT=no
+          BIZ_IMAGE_DIGEST=""
+          # Each caller is told what ITS absent image means, because the three
+          # situations have three different fixes.
+          case "$OA_DIGEST_MODE" in
+            deploy)
+              echo "[deploy] No image for this commit - the registry holds no $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA, so there is nothing main built for production to promote. Expected: the image the main pipeline of this same commit built and pushed (confirmation image-promotion 001). Fix: run a green pipeline on main over $CI_COMMIT_SHA first, then fast-forward production onto it - this deploy never builds a substitute." >&2
+              exit 1 ;;
+            build)
+              echo "[build] FATAL: $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA is not in the registry after this job did its work - either the push reported success and the registry does not hold it, or the tag this job found at its start is gone. Expected: the tag, since the digest of THIS build is what the production deploy will read back. Fix: re-run this job; if it repeats, the registry is accepting pushes it does not store." >&2
+              exit 1 ;;
+            *)
+              echo "[registry] $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA is not in the registry yet" ;;
+          esac ;;
+        *)
+          echo "[registry] FATAL: HEAD https://$CI_REGISTRY/v2/$CI_PROJECT_PATH/manifests/$CI_COMMIT_SHA answered HTTP $oa_status - expected 200 (the tag is there) or 404 (it is not). Fix: re-run once the registry answers one of those; this job never reads any other answer as an absent image." >&2
+          exit 1 ;;
+      esac
+      # end oa-registry-digest
+
 # The uniform of THIS commit, on every pipeline - confirmation
 # biz-service-manifest 010. Entry 008 places the BINDING run before the SSH step
 # of a production deploy; it never said that is the only place it runs, and read
@@ -121,39 +243,68 @@ validate-uniform:
     - if: $CI_COMMIT_BRANCH == "devel"
     - if: $CI_COMMIT_BRANCH == "production"
 
+# The image of a commit is built HERE and nowhere else (confirmation
+# image-promotion 001 point 1). production does not build: measured on
+# biz-pdfgen 59b2d59 on 2026-09-19, both branches pushed the same tag :<sha> and
+# the second push overwrote the first, so production ran sha256:1b2a136f… while
+# what main had tested was sha256:5f62e220… and was by then reachable under no
+# tag at all.
 build:
   stage: build
   image: docker:24
   services:
     - docker:24-dind
+  variables:
+    # A tag already in the registry is the ordinary case here (a re-run of this
+    # pipeline), never a defect - so the query below answers and the job decides.
+    # The SECOND question this job asks, after it has built or reused, is asked
+    # in mode "build", where an absent tag is a defect.
+    OA_DIGEST_MODE: "optional"
+  before_script:
+    # docker:24 is alpine and carries neither: the registry query needs both, and
+    # says so by name if they are missing.
+    - apk add --no-cache curl jq
   script:
-    # A password on a command line stands in the runner's process table for the
-    # length of the build and in every trace that echoes the command. The deploy
-    # host's own login has read it from stdin since it was written; both ends of
-    # one login now agree.
-    - echo "$CI_REGISTRY_PASSWORD" | docker login -u "$CI_REGISTRY_USER" --password-stdin "$CI_REGISTRY"
-    - docker build --target production -t $IMAGE -t $IMAGE_LATEST .
-    - docker push $IMAGE
-    - docker push $IMAGE_LATEST
-    # R1: the deploy target is the digest, and it exists only after the push.
+    - !reference [.oa-registry-digest, script]
     - |
-      # --format is not honoured by the buildx on the runner: it prints the whole
-      # descriptor (Name/MediaType/Digest), so the digest would be captured
-      # multi-line and build.env rejected as an invalid dotenv artifact. Parse
-      # the Digest field.
-      BIZ_IMAGE_DIGEST=$(docker buildx imagetools inspect "$IMAGE" | awk '/^Digest:/ { print $2; exit }')
-      if ! printf '%s' "$BIZ_IMAGE_DIGEST" | grep -qE '^sha256:[0-9a-f]{64}$'; then
-        echo "[build] FATAL: no valid digest resolved for $IMAGE - got '$BIZ_IMAGE_DIGEST', expected sha256:<64 hex>. The deploy would have no immutable target. Fix: confirm the push succeeded and that buildx can read $CI_REGISTRY." >&2
-        exit 1
+      if [ "$OA_IMAGE_PRESENT" = "yes" ]; then
+        # :<sha> is written once and never overwritten (image-promotion 001): a
+        # re-run reuses the image this commit already has, and leaves :latest
+        # where it is - re-running an older pipeline must not drag the moving tag
+        # backwards. Said out loud, because a step that decided not to work must
+        # not read like one that worked (automation-gates.md §5).
+        echo "[build] tag :$CI_COMMIT_SHA is already in the registry - not building, digest $BIZ_IMAGE_DIGEST"
+      else
+        # A password on a command line stands in the runner's process table for
+        # the length of the build and in every trace that echoes the command. The
+        # deploy host's own login has read it from stdin since it was written;
+        # both ends of one login now agree.
+        echo "$CI_REGISTRY_PASSWORD" | docker login -u "$CI_REGISTRY_USER" --password-stdin "$CI_REGISTRY"
+        docker build --target production -t $IMAGE -t $IMAGE_LATEST .
+        docker push $IMAGE
+        docker push $IMAGE_LATEST
       fi
-      echo "[build] $IMAGE -> $BIZ_IMAGE_DIGEST"
-      echo "BIZ_IMAGE_DIGEST=$BIZ_IMAGE_DIGEST" > build.env
+      # R1: the deploy target is the digest, and BOTH branches now read it from
+      # the registry, with the query above - the one the production deploy will
+      # ask too. `docker buildx imagetools inspect` used to answer it here, which
+      # made two implementations of one question ("what digest does :<sha>
+      # have") in one file (change-discipline.md § One rail per concern). What it
+      # reported was the same manifest digest, because this job builds one
+      # platform (a plain `docker build`, no --platform and no `buildx build
+      # --push`, so no index and no attestations) - but the day those differed,
+      # the pipeline would have proved one digest and promoted another.
+      OA_DIGEST_MODE=build
+    - !reference [.oa-registry-digest, script]
+    # The rest of THIS pipeline sees one digest, whether the image was built now
+    # or written by an earlier run. The production deploy does not read this
+    # artefact - it cannot, being in another pipeline - and asks the registry the
+    # same question instead.
+    - echo "BIZ_IMAGE_DIGEST=$BIZ_IMAGE_DIGEST" > build.env
   artifacts:
     reports:
       dotenv: build.env
   rules:
     - if: $CI_COMMIT_BRANCH == "main"
-    - if: $CI_COMMIT_BRANCH == "production"
 
 secret_detection:
   stage: secret-detection
@@ -177,6 +328,24 @@ deploy-production:
     # and the name is the service's registry identity — the one
     # api/config/services.json carries and the contract files its target under.
     DEPLOY_GATE_TARGETS: "biz:__REGISTRY_NAME__"
+    # A production deploy promotes the image main built for this same commit, so
+    # a commit with none is a deploy that must not happen (confirmation
+    # image-promotion 001 point 1). The query below says so by name and stops.
+    OA_DIGEST_MODE: "deploy"
+  # Nothing this job needs comes from another job. It used to take the digest
+  # from build's dotenv artefact, which is why the two jobs had to be in ONE
+  # pipeline - and on production there is no build job at all (image-promotion
+  # 001). Empty rather than absent: absent means "every artefact of every earlier
+  # stage", and a dotenv from somewhere else could then define BIZ_IMAGE_DIGEST
+  # behind this job's back (architecture-principles.md §8).
+  #
+  # It sits HERE, between two mappings, and not behind the before_script below:
+  # a comment indented under a key is not something the extractor of
+  # api/tests/scripts/biz-deploy-secret-handling.bats can end a literal block on,
+  # so this prose used to be handed to `sh` as part of the script (measured
+  # 2026-09-19 on b94c78e5, two suites red). templateImagePromotion.test.js keeps
+  # that shape now.
+  dependencies: []
   before_script:
     # Preconditions first, before anything is installed or deployed
     # (automation-gates.md §1 requirement 4). The post-deploy gate reads the
@@ -250,8 +419,6 @@ deploy-production:
         echo "  StrictHostKeyChecking yes"
         echo "  UserKnownHostsFile ~/.ssh/known_hosts"
       } > ~/.ssh/config
-  dependencies:
-    - build
   script:
     # The uniform of this checkout, complete, BEFORE anything is deployed
     # (confirmation biz-service-manifest 006 point 3, refined by 008): deploy is
@@ -260,12 +427,11 @@ deploy-production:
     # variable that turns this off (automation-gates.md §1 requirement 5).
     - npm ci
     - sh node_modules/@onlineapps/conn-orch-validator/templates/business-service/scripts/verify-deploy-uniform.sh "$CI_PROJECT_DIR" "https://gitlab-ci-token:${CI_JOB_TOKEN}@gitlab.com/${API_PROJECT_PATH}.git" "$API_UNIFORM_REF_DEPLOY"
-    # R1: refuse to deploy without the immutable target the build stage published.
-    - |
-      if [ -z "${BIZ_IMAGE_DIGEST:-}" ]; then
-        echo "[deploy] FATAL: BIZ_IMAGE_DIGEST is empty - the build stage publishes it as a dotenv artifact. Refusing to deploy without an immutable image target." >&2
-        exit 1
-      fi
+    # R1: the immutable target, resolved from the registry for THIS commit. The
+    # digest main proved is the digest production runs, and a commit main never
+    # built stops here with a named message rather than being built now
+    # (image-promotion 001; architecture-principles.md §3 No Fallbacks).
+    - !reference [.oa-registry-digest, script]
     - echo "Deploying $BIZ_IMAGE_DIGEST to $DEPLOY_HOST ($CI_ENVIRONMENT_URL)"
     # The read-only api clone the uniform step already made beside this checkout.
     # Derived from CI_PROJECT_DIR rather than spelled out a second time: the

package/templates/business-service/README.md

@@ -58,8 +58,10 @@ cp config/env-templates/service.env config/env-active/__SERVICE_NAME__.env
 docker compose up -d --build
 
 # 4. Start (production) — the image is pinned by digest, so it is PULLED, never
-#    built here: `docker build` refuses a tag that carries a digest. CI publishes
-#    the value; a manual run has to name the digest it wants.
+#    built here: `docker build` refuses a tag that carries a digest. The image is
+#    built once, by the main pipeline of that commit, and the production deploy
+#    reads its digest back from the registry (docs/80-setup/INSTALL.md §
+#    Production); a manual run has to name the digest it wants.
 BIZ_IMAGE_DIGEST=sha256:… docker compose -f docker-compose.production.yml pull
 BIZ_IMAGE_DIGEST=sha256:… docker compose -f docker-compose.production.yml up -d
 ```

package/templates/business-service/docs/80-setup/INSTALL.md

@@ -90,9 +90,37 @@ BIZ_IMAGE_DIGEST=sha256:… docker compose -f docker-compose.production.yml pull
 BIZ_IMAGE_DIGEST=sha256:… docker compose -f docker-compose.production.yml up -d
 ```
 
-CI publishes that digest from the build stage and the deploy job refuses to run
-without it (requirement R1). A first deploy also clones the checkout on the box
-if it is missing, so onboarding a new service needs no manual step there.
+### One image per commit: built on `main`, promoted to production
+
+The image of a commit is built **once**, by the `build` job of that commit's
+`main` pipeline, and pushed as `:<commit sha>`. That tag is written once and
+never overwritten: a re-run of the same pipeline finds it in the registry, says
+so, and neither builds nor pushes.
+
+A production deploy does not build. It asks the registry for the digest of
+`:<commit sha>` and deploys that — the image `main` tested, by digest
+(requirement R1). A commit whose `main` pipeline never produced one cannot reach
+production: the deploy stops with a message naming the commit, and never builds a
+substitute. The way forward is always the same one — a green `main` pipeline over
+that commit, then a fast-forward of `production` onto it. Owner decision:
+`api/docs/governance/confirmations/image-promotion.md` 001.
+
+Two things this rests on, and both are preconditions rather than details:
+
+- **Registry retention.** The image must still be there when `production`
+  deploys, and when a rollback needs the last green one. A cleanup policy on a
+  biz project may therefore be enabled only together with a rule that never
+  deletes a tag matching a 40-character commit sha; `:latest` and any other
+  moving tag are not part of this and may be cleaned freely.
+- **Rollback reads the same place.** There is no automatic rollback for the first
+  deployments (`api/docs/governance/confirmations/biz-rollback-first-deploy.md`
+  001): the return is the manual runbook step above — `docker compose up` with
+  the digest of the last green commit, read from the registry under that commit's
+  own `:<sha>`. Nothing writes that digest down anywhere else, which is exactly
+  why the retention rule is not optional.
+
+A first deploy also clones the checkout on the box if it is missing, so
+onboarding a new service needs no manual step there.
 
 ### Migrations at deploy time
 

package/templates/business-service/docs/80-setup/PLATFORM_MATRIX.md

@@ -54,8 +54,11 @@ the service.
   `@onlineapps/*` pins against the platform library set) and the service's
   dependencies installed — the test-coverage half asks the service's own jest
   which files it would run.
-- The build stage publishes the image digest as a dotenv artifact and the deploy
-  stage refuses to run without it; both stages are declared in `.gitlab-ci.yml`.
+- The build stage runs on `main` and builds the image of that commit once; the
+  deploy stage reads the digest of `:<commit sha>` back from the registry and
+  refuses to deploy a commit that has none. Both stages are declared in
+  `.gitlab-ci.yml`, and what each may do is
+  [INSTALL.md § Production](INSTALL.md).
 - Windows is outside the supported matrix.
 
 ## Status