[BUG]vm_workspace_diagnosis

Status Open
Maintainer reply None cached
Activity 0 comments · opened Aug 18, 2026

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

Claude 桌面应用更新后工作区无法启动 — 诊断报告

报告日期: 2026-08-18
应用: Claude 桌面应用(Claude-3p)
现象: 2026-08-17 应用更新后,工作区(VM)一直停留在"正在启动",无法使用。

结论(TL;DR)

更新后应用强制要求新版虚拟机镜像(bundle 版本 2a762adf),但该版本镜像在用户机器上无法正常启动:guest 系统内核执行到 Run /init 后即卡死,核心服务 coworkd 从未启动,host 等待 60 秒后报 VM connection timeout after 60 seconds。旧版镜像(6d1538ba)此前可正常启动。疑似新版镜像与当前环境不兼容或镜像自身有缺陷。

关键日志证据

日志位置:~/Library/Logs/Claude-3p/cowork_vm_node.logcoworkd.log

1. 更新后强制要求新版镜像,删除旧版

2026-08-17 22:45:03 [VM:start] Bundle version: 6d1538ba6fecc4e5c5583993c4b30bb1875f0f5a
2026-08-17 22:48:45 [error] [VM:steps] download_and_sdk_prepare failed (222569ms): Download failed.
2026-08-17 22:48:50 [deleteVMBundle] Deleted rootfs.img / vmlinuz / initrd ...
2026-08-18 09:56:13 [Bundle:status] rootfs.img version mismatch: have=6d1538ba..., want=2a762adfc2eea13eca0d113a2e6452ff00ae3f62

2. 新版镜像下载经历过超时(网络问题,已解决)

2026-08-18 09:58:53 [error] Startup failed: qun [VMDownloadError]: Request error: net::ERR_CONNECTION_TIMED_OUT
2026-08-18 10:12:38 [error] Startup failed: qun [VMDownloadError]: Download stalled: no data received for 30000ms

(多次重试后镜像文件最终下载/解压齐全。)

3. 新版镜像启动后 guest 卡死(核心故障)

host 侧每次启动约 60 秒后超时:

2026-08-18 15:50:09 [VM:start] Bundle version: 2a762adfc2eea13eca0d113a2e6452ff00ae3f62
2026-08-18 15:50:10 [VM:steps] vm_boot completed (374ms)
2026-08-18 15:50:10 [VM:steps] guest_vsock_connect started
2026-08-18 15:51:14 [error] [VM:start] Startup failed: VM connection timeout after 60 seconds

guest 侧(coworkd.log)只有内核启动记录,coworkd 服务从未启动:

[KERNEL] rtc_cmos 00:01: setting system clock to 2026-08-18T07:32:40 UTC
[KERNEL] Run /init as init process
(此后无任何 coworkd 日志)

4. 对照:旧版镜像此前正常启动

2026-08-18 01:52:40 [coworkd] starting coworkd
2026-08-18 01:52:44 [rpc] connecting to host CID=2 port=51234
2026-08-18 01:52:44 [rpc] connected successfully

失败启动时间线(2026-08-18,北京时间)

| 时间 | 结果 |
|---|---|
| 09:52 | 旧版镜像启动成功(之后应用 09:55 退出、09:56 删除旧镜像) |
| 11:16 | 新版镜像启动失败(guest 卡死) |
| 11:52 | 新版镜像启动失败(guest 卡死) |
| 12:05 | 新版镜像启动失败(guest 卡死) |
| 15:29 | 新版镜像启动失败(guest 卡死) |
| 15:32 | 新版镜像启动失败(guest 卡死,host 报 Connection timeout) |
| 15:50 | 新版镜像启动失败(guest 卡死,host 报 Connection timeout) |

建议官方排查方向

  1. 新版 bundle 2a762adf 的 rootfs/initrd 在部分机器上 init 阶段挂起,可能与硬件/虚拟化配置相关(用户为 Apple Silicon Mac)。
  2. 检查新版镜像的 initramfs 启动流程与 gVisor/vz 虚拟化的兼容性。
  3. 建议提供回退机制或旧版 bundle 下载入口。

附加信息

  • 磁盘空间充足(可用 67Gi)。
  • 用户处于中国大陆网络环境,新版镜像下载曾多次超时(ERR_CONNECTION_TIMED_OUT),但最终下载完整;此为次要问题,非根因。

What Should Happen?

vm_workspace_diagnosis

Error Messages/Logs

Steps to Reproduce

vm_workspace_diagnosis.md

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

Claude 1.32352.1 (6c6aa5) 2026-08-18T00:07:03.000Z

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

vm_workspace_diagnosis.md

View original on GitHub ↗