Python 3.9版Google Cloud Function突发构建失败求助

问题描述 #

我有一个基于Python 3.9的Google Cloud Function，此前运行完全正常，近期未对函数或依赖做任何修改，但手动测试时突然构建失败，报错找不到指定的builder镜像。

错误日志 #

[10:37:08 PM] - Preparing function sources ...
[10:37:08 PM] - Building function sources ...
[10:37:09 PM] - Error: Command failed: cd /gcf-working-dir/function-sources && pack build gcf-ui-function --builder gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder --pull-policy if-not-present --env GOOGLE_FUNCTION_TARGET=hello_http --env GOOGLE_FUNCTION_SIGNATURE_TYPE=cloudevent --env GOOGLE_DEVMODE=1 --env GOOGLE_RUNTIME_VERSION=3.9
ERROR: failed to build: failed to fetch builder image 'gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder:latest': image 'gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder:latest' does not exist on the daemon: not found

> at checkExecSyncError (node:child_process:826:11)
    at execSync (node:child_process:900:15)
    at buildFunction (/app/app.runfiles/google3/cloud/console/web/functions/external/testing_server/workers/build_function_worker.js:17:38)
    at Object.<anonymous> (/app/app.runfiles/google3/cloud/console/web/functions/external/testing_server/workers/build_function_worker.js:35:131)
    at Module._compile (node:internal/modules/cjs/loader:1095:14)
    at Object.Module._extensions..js (node:internal/modules/cjs/loader:1124:10)
    at Module.load (node:internal/modules/cjs/loader:975:32)
    at Function.Module._load (node:internal/modules/cjs/loader:816:12)
    at Function.executeUserEntryPoint [as runMain] (node:internal/modules/run_main:79:12)
    at MessagePort.<anonymous> (node:internal/main/worker_thread:183:24)
ERROR: failed to build: failed to fetch builder image 'gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder:latest': image 'gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder:latest' does not exist on the daemon: not found

[10:37:09 PM] - Error while building function sources. Please check logs.

已尝试的操作 #

• 反复确认Cloud Function的运行时设置为Python 3.9
• 验证gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder:latest镜像的可用性
• 多次重新部署函数，问题依旧

额外信息 #

• 函数此前无任何运行问题
• 近期未修改函数代码或依赖项
• Google Cloud SDK已更新至最新版本

可能的解决方法 #

1. 强制拉取最新镜像，跳过本地缓存

部署时修改构建的拉取策略，强制从仓库拉取最新镜像，避免本地缓存的过期镜像导致问题。使用gcloud命令部署时可指定参数：

gcloud functions deploy YOUR_FUNCTION_NAME \
  --runtime python39 \
  --trigger-http \
  --build-args="pull-policy=always"

2. 指定具体的镜像版本标签而非latest

latest标签可能被移除或更新，可通过GCR仓库页面查看该镜像的可用版本标签，替换命令中的latest为对应版本号，示例：

pack build gcf-ui-function \
  --builder gcr.io/gae-runtimes/buildpacks/google-gae-18/python/builder:20240501 \
  --pull-policy always \
  --env GOOGLE_FUNCTION_TARGET=hello_http \
  --env GOOGLE_FUNCTION_SIGNATURE_TYPE=cloudevent \
  --env GOOGLE_DEVMODE=1 \
  --env GOOGLE_RUNTIME_VERSION=3.9

3. 切换到GCP官方通用buildpack builder

如果特定的gae-18 builder已被废弃，尝试使用GCP通用的buildpack builder镜像：

pack build gcf-ui-function \
  --builder gcr.io/buildpacks/builder:v1 \
  --pull-policy always \
  --env GOOGLE_FUNCTION_TARGET=hello_http \
  --env GOOGLE_FUNCTION_SIGNATURE_TYPE=cloudevent \
  --env GOOGLE_DEVMODE=1 \
  --env GOOGLE_RUNTIME_VERSION=3.9

4. 清理构建缓存后重新部署

部署时添加--no-cache参数，强制跳过构建缓存，重新完成全量构建：

gcloud functions deploy YOUR_FUNCTION_NAME \
  --runtime python39 \
  --trigger-http \
  --no-cache

5. 检查项目服务账号的镜像拉取权限

确认Cloud Functions默认使用的服务账号（PROJECT_ID@appspot.gserviceaccount.com）拥有storage.objectViewer权限，确保能正常访问GCR仓库中的镜像。可通过IAM页面为该账号添加对应权限。

内容的提问来源于stack exchange，提问作者Veeee