Skip to content

Windows: 'Reset local data' fails when .openhuman directory is locked by another process #115

@ElioNeto

Description

@ElioNeto

Issue imported from tinyhumansai/openhuman#1615
Created at: unknown


Summary

On Windows, the config_reset_local_data RPC method fails when the .openhuman directory is locked by another process (os error 32 = ERROR_SHARING_VIOLATION).

Sentry Issues

Error: Failed to remove C:\Users\<username>\.openhuman: 另一个程序正在使用此文件,进程无法访问。 (os error 32)
Environment: production
Platform: Windows (x86_64)

Analysis

  • Triggered when users try to reset their local data
  • The core sidecar itself (or another OpenHuman instance) likely holds open file handles in .openhuman/
  • On Windows, files cannot be deleted while any process has them open
  • Two different users hit this independently

Suggested Fix

  1. Close own handles first — before deleting, close any open DB connections, file handles, etc. in the .openhuman directory
  2. Schedule deletion — mark for deletion on next restart if in-use
  3. User guidance — show "Close all OpenHuman windows and try again" if deletion fails due to file locks

Priority: Low — 2 events, edge case during reset flow

Metadata

Metadata

Assignees

No one assigned

    Labels

    rust-coreCore Rust runtime in src/: CLI, core_server, shared infrastructure.sentry-traced-bugBug identified via Sentry triage

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions