[Bug Report] artifact distribution failed on all 1 nodes: rpc error: code = Unimplemented desc = unknown service cubelet.services.images.v1.Images

[xiongxz]

Summary

After directly executing the following new installation command after the release of 0.2.0 on the previously installed old version, an error occurred while building the mirroring:
curl -sL https://cnb.cool/CubeSandbox/CubeSandbox/-/git/raw/master/deploy/one-click/online-install.sh | MIRROR=cn bash

error log:
2026/05/08 03:05:08 artifact_status: FAILED
2026/05/08 03:05:08 artifact_sha256: 6fc05d2b756efbd4654c423bf2619921234808bbd59b5b9df643c85af03ddacf
2026/05/08 03:05:08 error: artifact distribution failed on all 1 nodes: rpc error: code = Unimplemented desc = unknown service cubelet.services.images.v1.Images
cubemastercli run fail: artifact distribution failed on all 1 nodes: rpc error: code = Unimplemented desc = unknown service cubelet.services.images.v1.Images

Environment

• CubeSandbox version / commit:
• Host OS and kernel version:
• KVM info (modinfo kvm):
• Deployment mode:
• Relevant component:

Steps to Reproduce

1. on the previously installed old version. bash
2. curl -sL https://cnb.cool/CubeSandbox/CubeSandbox/-/git/raw/master/deploy/one-click/online-install.sh | MIRROR=cn
3. cubemastercli tpl create-from-image --image cube-sandbox-cn.tencentcloudcr.com/cube-sandbox/sandbox-code:latest --writable-layer-size 1G --expose-port 49999 --expose-port 49983 --probe 49999

Expected Behavior

Actual Behavior

paste logs here

Additional Context

[fslongjin]

Please reproduce the issue, then immediately package the Cube-related directories under /data/log/into a tarball and send it to me. I need to check the logs to pinpoint the problem.

[xiongxz]

the log tar file is a bit large and the upload keeps failing, so I've sent it to your email. thanks

[kangcp]

我在aws上也遇到了相同的问题

[ls-ggg]

Thanks for the report!
This error typically occurs when the Cubelet metadata gets corrupted due to unexpected filesystem-level operations. The ⁠unknown service cubelet.services.images.v1.Images⁠ RPC error indicates Cubelet is running in an inconsistent state.
You can try reinitializing Cubelet to recover the environment:
`
Step 1 — Stop Cube and its dependencies:
/usr/local/services/cubetoolbox/scripts/one-click/down-with-deps.sh

Step 2 — Restart Cube and its dependencies:
/usr/local/services/cubetoolbox/scripts/one-click/up-with-deps.sh

Step 3 — Reinitialize Cubelet data:
cubecli unsafe init

`
After the restart, retry the ⁠cubemastercli tpl create-from-image⁠ command. Please let us know if the issue persists.

[yangjie1993]

@ls-ggg 我执行这个也是失败报类似的grpc的问题cubecli unsafe init init will destroy ALL of the resource, continue only if you confirm [y/n]: y
cubecli run fail: rpc error: code = Unimplemented desc = unknown service cubelet.services.cubebox.v1.CubeLet

[ls-ggg]

@yangjie1993 I've updated my previous comment — ⁠cubecli unsafe init⁠ actually requires Cubelet to be running (it communicates via gRPC). Could you give it another try?

[yangjie1993]

@yangjie1993我更新了之前的评论——“cubecli unsafe init”实际上需要 Cubelet 正在运行（它通过 gRPC 通信）。您能再试一次吗？

试了还是一样的，我得cubelet是在运行

[yangjie1993]

@yangjie1993我更新了之前的评论——“cubecli unsafe init”实际上需要 Cubelet 正在运行（它通过 gRPC 通信）。您能再试一次吗？