unslothai/unsloth · error · HTTPException

Memory cleanup failed. See server logs for details.

Error message

Memory cleanup failed. See server logs for details.

What it means

HTTP 500 raised when ExportBackend.cleanup_memory completes but returns success=False — the backend ran, decided it could not free resources, and reported failure. The detail points to server logs because the reason (e.g. a worker refusing to release a model) is logged by the backend, not returned.

Source

Thrown at studio/backend/routes/export.py:131

        if isinstance(e, SidecarSwapInProgress):
            # Expected loss of the race against a sidecar install: retryable 409.
            raise HTTPException(status_code = 409, detail = str(e))
        logger.error(f"Error loading checkpoint: {e}", exc_info = True)
        raise HTTPException(
            status_code = 500,
            detail = "Failed to load checkpoint",
        )


@router.post("/cleanup", response_model = ExportOperationResponse)
async def cleanup_export_memory(current_subject: str = Depends(get_current_subject)):
    """Cleanup export-related models from memory (ExportBackend.cleanup_memory)."""
    try:
        backend = get_export_backend()
        success = await asyncio.to_thread(backend.cleanup_memory)

        if not success:
            raise HTTPException(
                status_code = 500,
                detail = "Memory cleanup failed. See server logs for details.",
            )

        return ExportOperationResponse(
            success = True,
            message = "Memory cleanup completed successfully",
        )
    except HTTPException:
        raise
    except Exception as e:
        logger.error(f"Error during export memory cleanup: {e}", exc_info = True)
        raise HTTPException(
            status_code = 500,
            detail = "Failed to cleanup export memory",
        )

View on GitHub (pinned to 203007d190)

Solutions

1. Check server logs around the cleanup call for the backend's reason.
2. Cancel any active export first (POST /export/cancel), then retry cleanup.
3. If a worker subprocess is stuck, restart the backend process to force-release VRAM.

const st = await api.get('/export/status');
if (st.is_export_active) await api.post('/export/cancel');

try {
  await api.post('/export/cleanup');
} catch (e) {
  if (e.status === 500) log('cleanup refused; check server logs');
}

When it happens

Trigger: POST /export/cleanup while a model is pinned by an active export/training subprocess, or when the cleanup routine internally catches an error and returns False instead of raising.

Common situations: Trying to free memory mid-export; a stuck worker subprocess holding references; platform differences in the cleanup path.

Related errors

• Failed to cleanup export memory
• Failed to load checkpoint
• Failed to cancel export
• Failed to get export status
• Failed to get export logs

AI-assisted analysis of unslothai/unsloth@203007d190 (2026-08-15). Data as JSON: /api/errors/6268ec00f84a2235. Report an issue: GitHub.