diff --git a/.ci/jenkins/lib/test-matrix.yaml b/.ci/jenkins/lib/test-matrix.yaml
index 3d742289a..23611d6a2 100644
--- a/.ci/jenkins/lib/test-matrix.yaml
+++ b/.ci/jenkins/lib/test-matrix.yaml
@@ -38,6 +38,62 @@ env:
   UCX_VERSION: v1.19.x
 
 steps:
+  - name: Pre-flight Check
+    parallel: false
+    run: |
+      # Check for hanging Docker containers
+      HANGING_CONTAINERS=$(docker ps -q --filter "name=nixl-ci-test")
+      if [ -n "$HANGING_CONTAINERS" ]; then
+        echo "ERROR: Found hanging Docker containers:"
+        docker ps --filter "name=nixl-ci-test"
+        exit 1
+      fi
+      echo "Pre-flight check: No hanging containers found"
+
+      # Initial workspace state
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=preflight"
+
+  - name: File Count Debug
+    parallel: false
+    run: |
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=initial"
+
+      # Fail if required files are missing
+      if [ "$GITLAB_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab directory is missing!"
+        ls -la ${WORKSPACE}
+        exit 1
+      fi
+      if [ "$BUILD_SH_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab/build.sh is missing!"
+        ls -la ${WORKSPACE}/.gitlab
+        exit 1
+      fi
+
+  - name: Stash Marker Debug
+    parallel: false
+    run: |
+      STASH_MARKER=$(ls -1 ${WORKSPACE} | grep -c '^scm-repo.tar$' || echo 0)
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      echo "DEBUG_FILES: count=$FILE_COUNT stash_marker=$STASH_MARKER stage=after_initial"
+
+      # Stash marker status explanation:
+      # stash_marker=0: Normal workspace (no tarball) - files extracted properly
+      # stash_marker=1: Tarball present - possible unstash failure, workspace not extracted
+      if [ "$STASH_MARKER" -eq 1 ]; then
+        echo "WARNING: scm-repo.tar found - possible Jenkins unstash failure"
+        echo "Workspace contents:"
+        ls -la ${WORKSPACE}
+      else
+        echo "INFO: No stash tarball found - workspace appears normal"
+      fi
+
   - name: Get Environment Info
     parallel: false
     run: |
@@ -61,16 +117,83 @@ steps:
       ibv_devinfo
       #ib_write_bw
 
+  - name: Start File Monitoring
+    parallel: false
+    run: |
+      # Start background file monitoring for deletions (after workspace is stable)
+      if command -v inotifywait >/dev/null 2>&1; then
+        nohup inotifywait -m -r -e delete,delete_self,moved_from \
+          --format '%T DELETED: %w%f (event: %e)' --timefmt '%Y-%m-%d %H:%M:%S' \
+          ${WORKSPACE}/ > /tmp/file-monitor-${BUILD_ID}-${axis_index}.log 2>&1 &
+        echo $! > /tmp/monitor-pid-${BUILD_ID}-${axis_index}
+        echo "File monitoring started, PID: $(cat /tmp/monitor-pid-${BUILD_ID}-${axis_index})"
+      else
+        echo "inotifywait not available - skipping file monitoring"
+        touch /tmp/file-monitor-${BUILD_ID}-${axis_index}.log
+      fi
+
+  - name: Post Environment Info Debug
+    parallel: false
+    run: |
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=post_env"
+
+      # Fail if required files are missing
+      if [ "$GITLAB_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab directory disappeared after environment info!"
+        exit 1
+      fi
+      if [ "$BUILD_SH_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab/build.sh disappeared after environment info!"
+        exit 1
+      fi
 
   - name: Build GPU Test Environment
     parallel: false
     run: |
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=pre_docker_build"
+
       docker build -t "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" -f .ci/dockerfiles/Dockerfile.gpu_test --build-arg BASE_IMAGE=${image} .
-    onfail: docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=post_docker_build"
+
+      # Fail if required files are missing
+      if [ "$GITLAB_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab directory disappeared after Docker build!"
+        exit 1
+      fi
+      if [ "$BUILD_SH_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab/build.sh disappeared after Docker build!"
+        exit 1
+      fi
+    onfail: |
+      # Show what files were deleted
+      if [ -f /tmp/file-monitor-${BUILD_ID}-${axis_index}.log ]; then
+        echo "=== FILES DELETED DURING BUILD ==="
+        cat /tmp/file-monitor-${BUILD_ID}-${axis_index}.log
+        echo "================================="
+      fi
+      docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+      # Comprehensive Docker cleanup
+      chmod +x .ci/scripts/docker-cleanup.sh
+      .ci/scripts/docker-cleanup.sh "${BUILD_ID}" "${axis_index}"
 
   - name: Run GPU Test Environment
     parallel: false
     run: |
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=pre_docker_run"
+
       docker run -dt --name "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" \
         --ulimit memlock=-1:-1 \
         --network=host \
@@ -82,33 +205,110 @@ steps:
         -v ${WORKSPACE}:${WORKSPACE} \
         -w ${WORKSPACE} \
         "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=post_docker_run"
+
+      # Fail if required files are missing
+      if [ "$GITLAB_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab directory disappeared after Docker run!"
+        exit 1
+      fi
+      if [ "$BUILD_SH_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab/build.sh disappeared after Docker run!"
+        exit 1
+      fi
     onfail: |
+      # Show what files were deleted
+      if [ -f /tmp/file-monitor-${BUILD_ID}-${axis_index}.log ]; then
+        echo "=== FILES DELETED DURING BUILD ==="
+        cat /tmp/file-monitor-${BUILD_ID}-${axis_index}.log
+        echo "================================="
+      fi
       docker rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
       docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+      # Comprehensive Docker cleanup
+      chmod +x .ci/scripts/docker-cleanup.sh
+      .ci/scripts/docker-cleanup.sh "${BUILD_ID}" "${axis_index}"
 
   - name: Build
     parallel: false
     run: |
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      GITLAB_EXISTS=$(test -d "${WORKSPACE}/.gitlab" && echo "YES" || echo "NO")
+      BUILD_SH_EXISTS=$(test -f "${WORKSPACE}/.gitlab/build.sh" && echo "YES" || echo "NO")
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=$GITLAB_EXISTS build_sh=$BUILD_SH_EXISTS stage=pre_build"
+
+      # Final check before attempting build
+      if [ "$GITLAB_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab directory is missing before build!"
+        exit 1
+      fi
+      if [ "$BUILD_SH_EXISTS" = "NO" ]; then
+        echo "ERROR: .gitlab/build.sh is missing before build!"
+        exit 1
+      fi
+
       docker exec -w ${WORKSPACE} -e UCX_VERSION=${UCX_VERSION} "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" /bin/bash -c ".gitlab/build.sh ${INSTALL_DIR}"
-    onfail: |
-      docker rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
-      docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
 
-  - name: Test CPP
-    parallel: false
-    run: |
-      docker exec -w ${WORKSPACE} "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" /bin/bash -c ".gitlab/test_cpp.sh ${INSTALL_DIR}"
-    onfail: |
+      FILE_COUNT=$(find ${WORKSPACE} -type f | wc -l)
+      echo "DEBUG_FILES: count=$FILE_COUNT gitlab_dir=YES build_sh=YES stage=post_build"
+    always: |
+      # Show what files were deleted
+      if [ -f /tmp/file-monitor-${BUILD_ID}-${axis_index}.log ]; then
+        echo "=== FILES DELETED DURING BUILD ==="
+        cat /tmp/file-monitor-${BUILD_ID}-${axis_index}.log
+        echo "================================="
+      fi
+      # Stop monitoring
+      if [ -f /tmp/monitor-pid-${BUILD_ID}-${axis_index} ]; then
+        kill $(cat /tmp/monitor-pid-${BUILD_ID}-${axis_index} 2>/dev/null) 2>/dev/null || true
+        rm -f /tmp/monitor-pid-${BUILD_ID}-${axis_index}
+      fi
       docker rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
       docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+      # Comprehensive Docker cleanup (on success or failure)
+      chmod +x .ci/scripts/docker-cleanup.sh
+      .ci/scripts/docker-cleanup.sh "${BUILD_ID}" "${axis_index}"
 
-  - name: Test Python
+  - name: Show File Monitoring Results
     parallel: false
     run: |
-      docker exec -w ${WORKSPACE} "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" /bin/bash -c ".gitlab/test_python.sh ${INSTALL_DIR}"
-    always: |
-      docker rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
-      docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+      echo "=== COMPLETE FILE MONITORING LOG ==="
+      if [ -f /tmp/file-monitor-${BUILD_ID}-${axis_index}.log ]; then
+        echo "Log file exists, size: $(wc -l < /tmp/file-monitor-${BUILD_ID}-${axis_index}.log) lines"
+        if [ -s /tmp/file-monitor-${BUILD_ID}-${axis_index}.log ]; then
+          echo "File deletions detected:"
+          cat /tmp/file-monitor-${BUILD_ID}-${axis_index}.log
+        else
+          echo "No file deletions detected during build"
+        fi
+      else
+        echo "Monitoring log file not found"
+      fi
+      echo "==================================="
+
+      # Final cleanup
+      rm -f /tmp/file-monitor-${BUILD_ID}-${axis_index}.log
+      rm -f /tmp/monitor-pid-${BUILD_ID}-${axis_index}
+
+# Tests commented out - problem occurs before tests
+#  - name: Test CPP
+#    parallel: false
+#    run: |
+#      docker exec -w ${WORKSPACE} "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" /bin/bash -c ".gitlab/test_cpp.sh ${INSTALL_DIR}"
+#    onfail: |
+#      docker rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+#      docker image rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
+
+#  - name: Test Python
+#    parallel: false
+#    run: |
+#      docker exec -w ${WORKSPACE} "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}" /bin/bash -c ".gitlab/test_python.sh ${INSTALL_DIR}"
+#    always: |
+#      docker rm -f "${JOB_BASE_NAME}-${BUILD_ID}-${axis_index}"
 
 # once this fix is merged we can use the following to stop/kill/rm the container instead of the cleanup command in each step
 # https://github.com/Mellanox/ci-demo/pull/111
diff --git a/.ci/scripts/docker-cleanup.sh b/.ci/scripts/docker-cleanup.sh
new file mode 100755
index 000000000..bec999874
--- /dev/null
+++ b/.ci/scripts/docker-cleanup.sh
@@ -0,0 +1,142 @@
+#!/bin/bash
+set -euo pipefail
+
+# Docker Cleanup Script - Comprehensive residue detection and cleanup
+# Usage: docker-cleanup.sh [BUILD_ID] [AXIS_INDEX]
+
+BUILD_ID="${1:-unknown}"
+AXIS_INDEX="${2:-unknown}"
+SCRIPT_NAME="docker-cleanup.sh"
+
+echo "=== $SCRIPT_NAME: Starting comprehensive Docker cleanup ==="
+echo "Build ID: $BUILD_ID, Axis: $AXIS_INDEX"
+echo "Timestamp: $(date '+%Y-%m-%d %H:%M:%S')"
+
+# Function to report findings
+report_finding() {
+    local category="$1"
+    local count="$2"
+    local details="$3"
+
+    if [ "$count" -gt 0 ]; then
+        echo "FOUND: $category - $count items"
+        echo "$details"
+        echo "---"
+    else
+        echo "CLEAN: $category - no residues found"
+    fi
+}
+
+# Function to safe cleanup with error handling
+safe_cleanup() {
+    local cmd="$1"
+    local desc="$2"
+
+    echo "Executing: $desc"
+    if eval "$cmd" 2>/dev/null; then
+        echo "SUCCESS: $desc"
+    else
+        echo "WARNING: Failed to execute: $desc (exit code: $?)"
+    fi
+}
+
+echo
+echo "=== PHASE 1: DETECTION ==="
+
+# 1. Check for nixl-ci-test containers (all states)
+echo "1. Checking for nixl-ci-test containers..."
+CONTAINERS=$(docker ps -a --filter "name=nixl-ci-test" --format "{{.ID}} {{.Names}} {{.Status}}" 2>/dev/null || echo "")
+CONTAINER_COUNT=$(echo "$CONTAINERS" | grep -c . || echo "0")
+report_finding "nixl-ci-test containers" "$CONTAINER_COUNT" "$CONTAINERS"
+
+# 2. Check for nixl-ci-test images
+echo "2. Checking for nixl-ci-test images..."
+IMAGES=$(docker images --filter "reference=nixl-ci-test*" --format "{{.Repository}}:{{.Tag}} {{.ID}} {{.Size}}" 2>/dev/null || echo "")
+IMAGE_COUNT=$(echo "$IMAGES" | grep -c . || echo "0")
+report_finding "nixl-ci-test images" "$IMAGE_COUNT" "$IMAGES"
+
+# 3. Check for dangling images
+echo "3. Checking for dangling images..."
+DANGLING=$(docker images -f "dangling=true" --format "{{.ID}} {{.CreatedAt}}" 2>/dev/null || echo "")
+DANGLING_COUNT=$(echo "$DANGLING" | grep -c . || echo "0")
+report_finding "dangling images" "$DANGLING_COUNT" "$DANGLING"
+
+# 4. Check for orphaned volumes
+echo "4. Checking for orphaned volumes..."
+VOLUMES=$(docker volume ls -f "dangling=true" --format "{{.Name}} {{.Driver}}" 2>/dev/null || echo "")
+VOLUME_COUNT=$(echo "$VOLUMES" | grep -c . || echo "0")
+report_finding "orphaned volumes" "$VOLUME_COUNT" "$VOLUMES"
+
+# 5. Check for dead containers
+echo "5. Checking for dead containers..."
+DEAD_CONTAINERS=$(docker ps -a --filter "status=dead" --format "{{.ID}} {{.Names}} {{.CreatedAt}}" 2>/dev/null || echo "")
+DEAD_COUNT=$(echo "$DEAD_CONTAINERS" | grep -c . || echo "0")
+report_finding "dead containers" "$DEAD_COUNT" "$DEAD_CONTAINERS"
+
+# 6. Check for exited containers (potential zombies)
+echo "6. Checking for old exited containers..."
+EXITED_CONTAINERS=$(docker ps -a --filter "status=exited" --filter "name=nixl-ci-test" --format "{{.ID}} {{.Names}} {{.Status}}" 2>/dev/null || echo "")
+EXITED_COUNT=$(echo "$EXITED_CONTAINERS" | grep -c . || echo "0")
+report_finding "exited nixl-ci-test containers" "$EXITED_COUNT" "$EXITED_CONTAINERS"
+
+# 7. Check Docker daemon health
+echo "7. Checking Docker daemon status..."
+DOCKER_INFO=$(docker info --format "{{.Containers}} containers, {{.Images}} images" 2>/dev/null || echo "DOCKER_DAEMON_ERROR")
+echo "Docker status: $DOCKER_INFO"
+
+# 8. Check for build cache
+echo "8. Checking Docker build cache..."
+BUILD_CACHE=$(docker system df --format "{{.Type}}: {{.Size}}" 2>/dev/null | grep -i cache || echo "")
+BUILD_CACHE_COUNT=$(echo "$BUILD_CACHE" | grep -c . || echo "0")
+report_finding "build cache" "$BUILD_CACHE_COUNT" "$BUILD_CACHE"
+
+echo
+echo "=== PHASE 2: CLEANUP ==="
+
+# Cleanup Phase 1: Specific nixl-ci-test resources
+if [ "$CONTAINER_COUNT" -gt 0 ]; then
+    echo "Cleaning up nixl-ci-test containers..."
+    safe_cleanup "docker ps -aq --filter 'name=nixl-ci-test' | xargs -r docker rm -f" "Remove all nixl-ci-test containers"
+fi
+
+if [ "$IMAGE_COUNT" -gt 0 ]; then
+    echo "Cleaning up nixl-ci-test images..."
+    safe_cleanup "docker images -q --filter 'reference=nixl-ci-test*' | xargs -r docker rmi -f" "Remove all nixl-ci-test images"
+fi
+
+# Cleanup Phase 2: General Docker cleanup
+if [ "$DEAD_COUNT" -gt 0 ]; then
+    echo "Cleaning up dead containers..."
+    safe_cleanup "docker container prune -f" "Remove dead containers"
+fi
+
+if [ "$DANGLING_COUNT" -gt 0 ]; then
+    echo "Cleaning up dangling images..."
+    safe_cleanup "docker image prune -f" "Remove dangling images"
+fi
+
+if [ "$VOLUME_COUNT" -gt 0 ]; then
+    echo "Cleaning up orphaned volumes..."
+    safe_cleanup "docker volume prune -f" "Remove orphaned volumes"
+fi
+
+# Cleanup Phase 3: Network cleanup (if needed)
+echo "Cleaning up unused networks..."
+safe_cleanup "docker network prune -f" "Remove unused networks"
+
+# Final verification
+echo
+echo "=== PHASE 3: VERIFICATION ==="
+echo "Final container count:"
+docker ps -a --filter "name=nixl-ci-test" --format "table {{.Names}}\t{{.Status}}" 2>/dev/null || echo "No containers found"
+
+echo "Final image count:"
+docker images --filter "reference=nixl-ci-test*" --format "table {{.Repository}}\t{{.Tag}}\t{{.Size}}" 2>/dev/null || echo "No images found"
+
+# System summary
+echo
+echo "=== SYSTEM SUMMARY ==="
+docker system df 2>/dev/null || echo "Could not get system usage"
+
+echo
+echo "=== $SCRIPT_NAME: Cleanup completed at $(date '+%Y-%m-%d %H:%M:%S') ==="