fix(ci): make the Doris 4.0.3 shared container healthy on CI runners

After bumping the shared CI container from doris-all-in-one-2.1.0 to 4.0.3-all-slim
(so CI matches the fixture image), the doris tests began timing out the 1h job cap.
Root cause, found by reproducing the container locally: 4.0.3's start_be.sh
hard-`exit 1`s unless vm.max_map_count >= 2000000 (it also gates on swap and
ulimit). The 2.1.0 image had no such gate, which is exactly why it passed in 13min
and 4.0.3 did not. On the runner's low default the BE exits on start; its death
ends the entrypoint's `wait $child_pid`, so the container exits 0 and never reports
healthy — forcing every test onto the slow per-test self-boot path (480s nextest
timeout x 4 retries).

- Raise vm.max_map_count to 2000000 on the runner before launch (inherited by all
  containers) and pass SKIP_CHECK_ULIMIT=true to bypass the swap/ulimit gates
  without swapoff'ing the runner. Verified locally: BE reaches alive=true.
- Self-boot fixture sets SKIP_CHECK_ULIMIT=true too, so the fallback path and
  local Linux dev aren't blocked by the same gates.
- Cap FE heap to 2GB and BE mem_limit to 4GB: the container boots beside the cargo
  build on a 16GB runner, and 4.0.3 otherwise sizes for a dedicated host (8GB FE
  heap + ~90%-of-RAM BE), so bounding it keeps room for rustc.
- Make the shared-container wait non-fatal and dump docker logs/inspect on the
  unhealthy path, so a launch failure degrades to self-boot instead of aborting
  the step.

Also rustfmt the sqlx AssertSqlSafe call sites introduced by the sqlx 0.9 migration
in the preceding master merge.

Verified in CI: shared container reports healthy and test-1/test-2 pass (~13min).

Author: ryankert01

.github/actions/rust/pre-merge/action.yml

@@ -229,9 +229,26 @@ runs:
         DORIS_STARTED=0
         if [[ "$RUNNER_OS" == "Linux" ]] && { [[ -z "$NEXTEST_FILTER" ]] || echo "$NEXTEST_FILTER" | grep -qi 'integration'; }; then
           echo "::group::Launch shared Doris container (boots during build)"
+          # ROOT CAUSE of the regression: 4.0.3's start_be.sh hard-`exit 1`s unless
+          # `vm.max_map_count >= 2000000` (the 2.1.0 image had no such gate, which is
+          # why it worked). The runner default is far lower, so the BE died on start;
+          # its death ends the entrypoint's `wait`, the container exits, and every
+          # doris test falls onto the slow per-test self-boot path → 1h timeout.
+          # Raise it on the host (inherited by all containers, shared and self-boot);
+          # reproduced locally that this brings the BE alive. `SKIP_CHECK_ULIMIT`
+          # additionally bypasses start_be.sh's swap/ulimit gates so we needn't
+          # swapoff the runner.
+          sudo sysctl -w vm.max_map_count=2000000 || true
+          # Secondary: this boots *concurrently with the cargo build* on a 16GB
+          # runner, and 4.0.3 sizes for a dedicated host (FE 8GB -Xms heap, BE
+          # mem_limit `auto` ≈ 90% of RAM). Cap both (FE 2GB + BE 4GB) so Doris fits
+          # beside the build, ~matching the 2.1.0 footprint. Override the entrypoint
+          # to rewrite the confs, then hand off to the image's normal entry_point.sh.
           if docker run -d --name iggy-doris-shared \
               -p 8030 -p 9030 -p 8040:8040 \
-              apache/doris:4.0.3-all-slim; then
+              -e SKIP_CHECK_ULIMIT=true \
+              --entrypoint bash apache/doris:4.0.3-all-slim \
+              -c "sed -i 's/-Xmx8192m -Xms8192m/-Xmx2048m -Xms2048m/' /opt/apache-doris/fe/conf/fe.conf && echo 'mem_limit = 4096M' >> /opt/apache-doris/be/conf/be.conf && exec bash /usr/local/bin/entry_point.sh"; then
             DORIS_STARTED=1
           else
             echo "::warning::Failed to start shared Doris container; doris tests will self-boot"
@@ -296,6 +313,18 @@ runs:
             echo "Shared Doris ready: FE ${DORIS_FE_URL}, MySQL ${MYSQL_PORT}"
           else
             echo "::warning::Shared Doris did not become healthy; doris tests will self-boot"
+            # Diagnostics: why did 4.0.3-all-slim not become healthy as a bare
+            # `docker run` (2.1.0 did, and 4.0.3 self-boots fine via testcontainers)?
+            # Capture container state (Exited? OOMKilled? exit code) and the tail
+            # of its logs so we can tell crash/OOM apart from slow-boot.
+            echo "::group::Shared Doris diagnostics"
+            docker ps -a --filter name=iggy-doris-shared \
+              --format 'status={{.Status}} state={{.State}}' 2>&1 || true
+            docker inspect iggy-doris-shared \
+              --format 'ExitCode={{.State.ExitCode}} OOMKilled={{.State.OOMKilled}} Error={{.State.Error}}' 2>&1 || true
+            echo "--- last 60 lines of container logs ---"
+            docker logs --tail 60 iggy-doris-shared 2>&1 || true
+            echo "::endgroup::"
             docker rm -f iggy-doris-shared >/dev/null 2>&1 || true
           fi
           echo "::endgroup::"

core/integration/tests/connectors/fixtures/doris/container.rs

@@ -212,6 +212,13 @@ impl DorisContainer {
                     .with_port(FE_HTTP_PORT.tcp())
                     .with_expected_status_code(200u16),
             ))
+            // Doris 4.0.3's start_be.sh refuses to start (exit 1) unless
+            // vm.max_map_count >= 2000000 and swap is off and ulimit -n >= 60000.
+            // Hosts (CI runners, dev laptops) rarely satisfy all three, and the BE
+            // dying makes the container exit before it ever reports healthy. Skip
+            // those gate checks; the test workload is tiny so the conservative
+            // thresholds don't matter. (CI also raises vm.max_map_count on the host.)
+            .with_env_var("SKIP_CHECK_ULIMIT", "true")
             .with_network(unique_network)
             .with_mapped_port(0, FE_HTTP_PORT.tcp())
             .with_mapped_port(0, FE_MYSQL_PORT.tcp())
@@ -357,12 +364,12 @@ impl DorisContainer {
             "CREATE DATABASE IF NOT EXISTS {}",
             self.database
         )))
-            .execute(&pool)
-            .await
-            .map_err(|e| TestBinaryError::FixtureSetup {
-                fixture_type: "DorisContainer".to_string(),
-                message: format!("Failed to create test database: {e}"),
-            })?;
+        .execute(&pool)
+        .await
+        .map_err(|e| TestBinaryError::FixtureSetup {
+            fixture_type: "DorisContainer".to_string(),
+            message: format!("Failed to create test database: {e}"),
+        })?;
         Ok(())
     }
 }
@@ -457,11 +464,11 @@ pub trait DorisOps: Sync {
         let row = sqlx::raw_sql(sqlx::AssertSqlSafe(format!(
             "SELECT COUNT(*) AS c FROM {database}.{table}"
         )))
-            .fetch_one(&pool)
-            .await
-            .map_err(|e| TestBinaryError::InvalidState {
-                message: format!("Failed to count rows in {database}.{table}: {e}"),
-            })?;
+        .fetch_one(&pool)
+        .await
+        .map_err(|e| TestBinaryError::InvalidState {
+            message: format!("Failed to count rows in {database}.{table}: {e}"),
+        })?;
         // Doris returns COUNT(*) as BIGINT; sqlx decodes that to i64.
         let count: i64 = row.try_get(0).map_err(|e| TestBinaryError::InvalidState {
             message: format!("Failed to read count column: {e}"),