启用Kaniko缓存后gcloud run deploy触发KeyError:'images'崩溃问题

镜像构建完成后执行gcloud run deploy命令前，gcloud CLI突然崩溃并报错：

2022-09-19T06:05:37.8347067Z Building Container....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................done
2022-09-19T06:05:37.8351347Z failed
2022-09-19T06:05:37.8352361Z Deployment failed
2022-09-19T06:05:37.8839076Z ERROR: gcloud crashed (KeyError): 'images'
2022-09-19T06:05:37.8839690Z 
2022-09-19T06:05:37.8840175Z If you would like to report this issue, please run the following command:
2022-09-19T06:05:37.8840690Z   gcloud feedback
2022-09-19T06:05:37.8840875Z 
2022-09-19T06:05:37.8841213Z To check gcloud for common problems, please run the following command:
2022-09-19T06:05:37.8847153Z   gcloud info --run-diagnostics

问题原因 #

执行部署脚本的同一用户，在同一构建机的另一个独立构建脚本中启用了Kaniko缓存：

# Enable Kaniko cache to speed up the build
gcloud config set builds/use_kaniko True
gcloud config set builds/kaniko_cache_ttl 672 # 4 weeks

这是Cloud SDK的已知bug，会导致gcloud run deploy --source与gcloud run deploy --image命令产生冲突，最终触发上述崩溃。

解决方法 #

启用Kaniko缓存后，同一用户同一机器上的所有gcloud run deploy命令必须使用--image参数。只需将原来的gcloud run deploy --source拆分为两个独立命令：

• gcloud builds submit --tag IMAGE_ID（替换IMAGE_ID为你的镜像标识）
• gcloud run deploy --image IMAGE_ID

内容的提问来源于stack exchange，提问作者Anton Komarov