[Fix] (CI): stabilize Android retrace jobs

This commit is contained in:
2026-07-18 19:52:04 -04:00
parent 626c7f26fd
commit fc0688c223
2 changed files with 85 additions and 32 deletions
+60 -32
View File
@@ -227,6 +227,8 @@ jobs:
env: env:
AVD_NAME: mobilegl-ci AVD_NAME: mobilegl-ci
ANDROID_AVD_HOME: ${{ github.workspace }}/.android/avd ANDROID_AVD_HOME: ${{ github.workspace }}/.android/avd
ANDROID_HOME: ${{ github.workspace }}/.android/sdk
ANDROID_SDK_ROOT: ${{ github.workspace }}/.android/sdk
steps: steps:
- name: Checkout repo - name: Checkout repo
uses: actions/checkout@v6 uses: actions/checkout@v6
@@ -245,11 +247,11 @@ jobs:
with: with:
path: | path: |
${{ env.ANDROID_AVD_HOME }} ${{ env.ANDROID_AVD_HOME }}
/usr/local/lib/android/sdk/emulator ${{ env.ANDROID_SDK_ROOT }}/emulator
/usr/local/lib/android/sdk/platform-tools ${{ env.ANDROID_SDK_ROOT }}/platform-tools
/usr/local/lib/android/sdk/platforms/android-35 ${{ env.ANDROID_SDK_ROOT }}/platforms/android-35
/usr/local/lib/android/sdk/system-images/android-35/google_apis/x86_64 ${{ env.ANDROID_SDK_ROOT }}/system-images/android-35/google_apis/x86_64
key: ${{ runner.os }}-mobilegl-avd-api35-google_apis-x86_64-pixel_6-v1-${{ hashFiles('android-plugin/run-avd-ci.sh') }} key: ${{ runner.os }}-mobilegl-avd-api35-google_apis-x86_64-pixel_6-v2-${{ hashFiles('android-plugin/run-avd-ci.sh') }}
- name: Create AVD - name: Create AVD
if: steps.android-avd-cache.outputs.cache-hit != 'true' if: steps.android-avd-cache.outputs.cache-hit != 'true'
@@ -274,6 +276,8 @@ jobs:
env: env:
AVD_NAME: mobilegl-ci AVD_NAME: mobilegl-ci
ANDROID_AVD_HOME: ${{ github.workspace }}/.android/avd ANDROID_AVD_HOME: ${{ github.workspace }}/.android/avd
ANDROID_HOME: ${{ github.workspace }}/.android/sdk
ANDROID_SDK_ROOT: ${{ github.workspace }}/.android/sdk
strategy: strategy:
fail-fast: false fail-fast: false
max-parallel: 4 max-parallel: 4
@@ -288,7 +292,7 @@ jobs:
- name: Set Swap Space - name: Set Swap Space
uses: pierotofy/set-swap-space@v1.0 uses: pierotofy/set-swap-space@v1.0
with: with:
swap-size-gb: 16 swap-size-gb: 8
- name: Checkout repo - name: Checkout repo
uses: actions/checkout@v6 uses: actions/checkout@v6
@@ -324,11 +328,11 @@ jobs:
with: with:
path: | path: |
${{ env.ANDROID_AVD_HOME }} ${{ env.ANDROID_AVD_HOME }}
/usr/local/lib/android/sdk/emulator ${{ env.ANDROID_SDK_ROOT }}/emulator
/usr/local/lib/android/sdk/platform-tools ${{ env.ANDROID_SDK_ROOT }}/platform-tools
/usr/local/lib/android/sdk/platforms/android-35 ${{ env.ANDROID_SDK_ROOT }}/platforms/android-35
/usr/local/lib/android/sdk/system-images/android-35/google_apis/x86_64 ${{ env.ANDROID_SDK_ROOT }}/system-images/android-35/google_apis/x86_64
key: ${{ runner.os }}-mobilegl-avd-api35-google_apis-x86_64-pixel_6-v1-${{ hashFiles('android-plugin/run-avd-ci.sh') }} key: ${{ runner.os }}-mobilegl-avd-api35-google_apis-x86_64-pixel_6-v2-${{ hashFiles('android-plugin/run-avd-ci.sh') }}
- name: Download retrace APK - name: Download retrace APK
uses: actions/download-artifact@v8 uses: actions/download-artifact@v8
@@ -378,27 +382,51 @@ jobs:
if [ "${{ matrix.case.coherent_as_flush || false }}" = "true" ]; then if [ "${{ matrix.case.coherent_as_flush || false }}" = "true" ]; then
extra_retrace_args+=(--coherent-as-flush) extra_retrace_args+=(--coherent-as-flush)
fi fi
timeout "$(( ${{ matrix.case.timeout_seconds }} + 300 ))" sh android-plugin/trace-replay-ci.sh \
--apk-file "${apk_file}" \ run_retrace() {
--package top.mobilegl.plugin.trace \ timeout "$(( ${{ matrix.case.timeout_seconds }} + 300 ))" sh android-plugin/trace-replay-ci.sh \
--backend "${{ matrix.backend.name }}" \ --apk-file "${apk_file}" \
--result-root android-retrace-result \ --package top.mobilegl.plugin.trace \
--fixture-root android-retrace-fixture \ --backend "${{ matrix.backend.name }}" \
--case "${{ matrix.case.name }}" \ --result-root android-retrace-result \
--trace-archive "${{ matrix.case.trace_archive }}" \ --fixture-root android-retrace-fixture \
--trace-file "${{ matrix.case.trace_file }}" \ --case "${{ matrix.case.name }}" \
--golden "${{ matrix.case.golden }}" \ --trace-archive "${{ matrix.case.trace_archive }}" \
--alternate-golden "${{ matrix.case.alternate_golden || '' }}" \ --trace-file "${{ matrix.case.trace_file }}" \
--target-call "${{ matrix.case.target_call }}" \ --golden "${{ matrix.case.golden }}" \
--width "${{ matrix.case.width }}" \ --alternate-golden "${{ matrix.case.alternate_golden || '' }}" \
--height "${{ matrix.case.height }}" \ --target-call "${{ matrix.case.target_call }}" \
--ssim-threshold "${{ matrix.case.ssim_threshold || '0.99' }}" \ --width "${{ matrix.case.width }}" \
--crop-x "${{ matrix.case.crop_x }}" \ --height "${{ matrix.case.height }}" \
--crop-y "${{ matrix.case.crop_y }}" \ --ssim-threshold "${{ matrix.case.ssim_threshold || '0.99' }}" \
--crop-width "${{ matrix.case.crop_width }}" \ --crop-x "${{ matrix.case.crop_x }}" \
--crop-height "${{ matrix.case.crop_height }}" \ --crop-y "${{ matrix.case.crop_y }}" \
--timeout-seconds "${{ matrix.case.timeout_seconds }}" \ --crop-width "${{ matrix.case.crop_width }}" \
"${extra_retrace_args[@]}" --crop-height "${{ matrix.case.crop_height }}" \
--timeout-seconds "${{ matrix.case.timeout_seconds }}" \
"${extra_retrace_args[@]}"
}
retrace_status=0
run_retrace || retrace_status=$?
if [ "${retrace_status}" -eq 75 ]; then
echo "::warning::Android emulator infrastructure failed; restarting it and retrying this retrace once."
sh android-plugin/run-avd-ci.sh stop \
--avd-name "${AVD_NAME}" \
--emulator-log "${EMULATOR_LOG}" \
--pid-file "${EMULATOR_PID_FILE}"
adb kill-server || true
sleep 2
sh android-plugin/run-avd-ci.sh start \
--avd-name "${AVD_NAME}" \
--gpu "${{ matrix.backend.gpu }}" \
--emulator-log "${EMULATOR_LOG}" \
--pid-file "${EMULATOR_PID_FILE}" \
--boot-timeout 300
run_retrace
elif [ "${retrace_status}" -ne 0 ]; then
exit "${retrace_status}"
fi
- name: Collect retrace summary inputs - name: Collect retrace summary inputs
if: always() if: always()
+25
View File
@@ -3,6 +3,7 @@ set -eu
ADB="${ADB:-adb}" ADB="${ADB:-adb}"
PYTHON="${PYTHON:-python3}" PYTHON="${PYTHON:-python3}"
INFRASTRUCTURE_FAILURE_EXIT_CODE=75
usage() { usage() {
cat <<'EOF' cat <<'EOF'
@@ -171,6 +172,12 @@ app_dir="/data/user/0/${package_name}/files/trace-replay"
collect_run_diagnostics() { collect_run_diagnostics() {
diagnostics_dir="$1" diagnostics_dir="$1"
mkdir -p "${diagnostics_dir}" mkdir -p "${diagnostics_dir}"
adb_state="$(adb_device_path get-state 2>/dev/null | tr -d '\r' || true)"
printf '%s\n' "${adb_state}" > "${diagnostics_dir}/adb-state.txt"
: > "${diagnostics_dir}/logcat.txt"
if [ "${adb_state}" != "device" ]; then
return
fi
"${ADB}" logcat -d -t 2000 > "${diagnostics_dir}/logcat.txt" || true "${ADB}" logcat -d -t 2000 > "${diagnostics_dir}/logcat.txt" || true
adb_device_path shell pidof "${package_name}" > "${diagnostics_dir}/pidof.txt" 2>&1 || true adb_device_path shell pidof "${package_name}" > "${diagnostics_dir}/pidof.txt" 2>&1 || true
adb_device_path shell dumpsys activity activities > "${diagnostics_dir}/activity.txt" 2>&1 || true adb_device_path shell dumpsys activity activities > "${diagnostics_dir}/activity.txt" 2>&1 || true
@@ -179,6 +186,20 @@ collect_run_diagnostics() {
adb_device_path exec-out run-as "${package_name}" cat "${app_dir}/output/mobilegl.log" > "${diagnostics_dir}/mobilegl.log" || true adb_device_path exec-out run-as "${package_name}" cat "${app_dir}/output/mobilegl.log" > "${diagnostics_dir}/mobilegl.log" || true
} }
is_infrastructure_failure() {
diagnostics_dir="$1"
adb_state="$(cat "${diagnostics_dir}/adb-state.txt" 2>/dev/null || true)"
if [ "${adb_state}" != "device" ]; then
echo "trace-replay-ci.sh: Android device is unavailable (state: ${adb_state:-unknown})" >&2
return 0
fi
if grep -Eq 'Fatal signal [0-9]+.*[(]system_server[)]|F system_server[ :]' "${diagnostics_dir}/logcat.txt"; then
echo "trace-replay-ci.sh: Android system_server crashed during retrace" >&2
return 0
fi
return 1
}
copy_app_artifact() { copy_app_artifact() {
source_path="$1" source_path="$1"
destination_path="$2" destination_path="$2"
@@ -310,6 +331,10 @@ run_retrace() {
grep -E 'MobileGLTraceRunner|AndroidRuntime|FATAL EXCEPTION|trace_replay|MobileGL|libc' "${result_dir}/logcat.txt" | tail -200 >&2 || true grep -E 'MobileGLTraceRunner|AndroidRuntime|FATAL EXCEPTION|trace_replay|MobileGL|libc' "${result_dir}/logcat.txt" | tail -200 >&2 || true
echo "trace-replay-ci.sh: app trace-replay files:" >&2 echo "trace-replay-ci.sh: app trace-replay files:" >&2
cat "${result_dir}/app-files.txt" >&2 || true cat "${result_dir}/app-files.txt" >&2 || true
if is_infrastructure_failure "${result_dir}"; then
echo "trace-replay-ci.sh: requesting one infrastructure retry" >&2
exit "${INFRASTRUCTURE_FAILURE_EXIT_CODE}"
fi
exit 1 exit 1
fi fi
adb_device_path exec-out run-as "${package_name}" cat "${app_dir}/output/result.json" > "${result_dir}/result.json" adb_device_path exec-out run-as "${package_name}" cat "${app_dir}/output/result.json" > "${result_dir}/result.json"