From caadd6ca792898f12f130d9b5fc418efa24c5392 Mon Sep 17 00:00:00 2001
From: Chris Lu <chrislusf@users.noreply.github.com>
Date: Wed, 10 Jun 2026 13:26:30 -0700
Subject: [PATCH] ci(s3tables): stop Lakekeeper flaking on Docker Hub pull
 timeouts (#9920)

* ci(s3tables): drop docker pre-pull from Lakekeeper job

The lakekeeper repro is pure Go against the local weed binary; the job
kept failing on Docker Hub timeouts pulling python:3 and localstack
images the test never runs. Also drop the stale python-in-docker
comments left from the old harness.

* ci(s3tables): serve python:3 from GHA cache in the STS job

Retried pulls still die when both mirror.gcr.io and registry-1.docker.io
are unreachable from the runner. Cache the saved image tarball under a
weekly key: an exact hit skips the registry entirely, a miss pulls fresh
and refreshes the cache, and a stale tarball from a previous week is the
fallback when Docker Hub is down.

* ci(spark): pre-pull the spark tag the test actually runs

The workflow warmed apache/spark:3.5.8 with retries while the
testcontainers setup runs apache/spark:3.5.1, so the real image was
pulled at test time with no retry at all.
---
 .github/workflows/s3-spark-tests.yml        |  2 +-
 .github/workflows/s3-tables-tests.yml       | 40 +++++++++++++--------
 test/s3tables/lakekeeper/lakekeeper_test.go |  2 --
 3 files changed, 27 insertions(+), 17 deletions(-)

diff --git a/.github/workflows/s3-spark-tests.yml b/.github/workflows/s3-spark-tests.yml
index 8154a323b..6214a68f8 100644
--- a/.github/workflows/s3-spark-tests.yml
+++ b/.github/workflows/s3-spark-tests.yml
@@ -45,7 +45,7 @@ jobs:
       - name: Pre-pull Spark image
         run: |
           pull() { for i in 1 2 3; do docker pull "$1" && return 0; sleep 15; done; return 1; }
-          pull apache/spark:3.5.8
+          pull apache/spark:3.5.1
 
       - name: Run S3 Spark integration tests
         working-directory: test/s3/spark
diff --git a/.github/workflows/s3-tables-tests.yml b/.github/workflows/s3-tables-tests.yml
index bbc28e412..2b4404eee 100644
--- a/.github/workflows/s3-tables-tests.yml
+++ b/.github/workflows/s3-tables-tests.yml
@@ -574,10 +574,34 @@ jobs:
           echo '{"registry-mirrors": ["https://mirror.gcr.io"]}' | sudo tee /etc/docker/daemon.json
           sudo systemctl restart docker
 
-      - name: Pre-pull Python image
+      - name: Week stamp for image cache key
+        id: week
+        run: echo "week=$(date -u +%G-%V)" >> "$GITHUB_OUTPUT"
+
+      - name: Restore python:3 image cache
+        id: python-image
+        uses: actions/cache@v5
+        with:
+          path: /tmp/python3-image.tar
+          key: python3-image-${{ steps.week.outputs.week }}
+          restore-keys: |
+            python3-image-
+
+      - name: Load or pull python:3
         run: |
+          if [ "${{ steps.python-image.outputs.cache-hit }}" = "true" ]; then
+            docker load -i /tmp/python3-image.tar
+            exit 0
+          fi
           pull() { for i in 1 2 3; do docker pull "$1" && return 0; sleep 15; done; return 1; }
-          pull python:3
+          if pull python:3; then
+            docker save -o /tmp/python3-image.tar python:3
+          elif [ -f /tmp/python3-image.tar ]; then
+            # Docker Hub unreachable; fall back to last week's cached image
+            docker load -i /tmp/python3-image.tar
+          else
+            exit 1
+          fi
 
       - name: Run go mod tidy
         run: go mod tidy
@@ -639,18 +663,6 @@ jobs:
           go-version-file: 'go.mod'
         id: go
 
-      - name: Configure Docker Hub mirror
-        run: |
-          echo '{"registry-mirrors": ["https://mirror.gcr.io"]}' | sudo tee /etc/docker/daemon.json
-          sudo systemctl restart docker
-
-      - name: Pre-pull images
-        run: |
-          pull() { for i in 1 2 3; do docker pull "$1" && return 0; sleep 15; done; return 1; }
-          pull python:3
-          # localstack is optional; best-effort pull
-          for i in 1 2 3; do docker pull localstack/localstack:latest && break; sleep 15; done
-
       - name: Run go mod tidy
         run: go mod tidy
 
diff --git a/test/s3tables/lakekeeper/lakekeeper_test.go b/test/s3tables/lakekeeper/lakekeeper_test.go
index 41059c4ce..7076cf187 100644
--- a/test/s3tables/lakekeeper/lakekeeper_test.go
+++ b/test/s3tables/lakekeeper/lakekeeper_test.go
@@ -68,7 +68,6 @@ func TestLakekeeperIntegration(t *testing.T) {
 	env.StartSeaweedFS(t)
 	fmt.Printf(">>> SeaweedFS started.\n")
 
-	// Run python script in docker to test STS and S3 operations
 	runLakekeeperRepro(t, env)
 }
 
@@ -80,7 +79,6 @@ func TestLakekeeperTableBucketIntegration(t *testing.T) {
 	env.StartSeaweedFS(t)
 	fmt.Printf(">>> SeaweedFS started.\n")
 
-	// Run python script in docker to test STS and S3 Tables operations
 	runLakekeeperTableBucketRepro(t, env)
 }