infiniflow/ragflow · error · RuntimeError

Failed to create sandbox workspace: {result.stderr_text or '

Error message

Failed to create sandbox workspace: {result.stderr_text or 'unknown error'}

What it means

Raised after the sandbox is created but the bootstrap 'mkdir -p <work_dir>/artifacts' command inside it exits non-zero. The remote workspace under /home/tenki could not be created, so the sandbox is terminated via _safe_terminate() and the create_instance() call fails.

Source

Thrown at agent/sandbox/providers/tenki.py:165

            raise RuntimeError(f"Tenki quota exceeded: {exc}") from exc
        except errors.RateLimitedError as exc:
            raise RuntimeError(f"Tenki rate limited, please retry: {exc}") from exc
        except errors.UnauthorizedError as exc:
            raise SandboxProviderConfigError("Tenki authentication failed: check the API key.") from exc
        except Exception as exc:
            # Satisfy the base contract: any other SDK failure becomes RuntimeError.
            raise RuntimeError(f"Failed to create Tenki sandbox: {exc}") from exc

        remote_work_dir = posixpath.join(SANDBOX_HOME, f"ragflow-codeexec-{uuid.uuid4().hex}")
        try:
            result = sandbox.exec(
                "mkdir",
                "-p",
                posixpath.join(remote_work_dir, "artifacts"),
                timeout=min(self.timeout, 10),
            )
            if result.exit_code != 0:
                raise RuntimeError(f"Failed to create sandbox workspace: {result.stderr_text or 'unknown error'}")
        except Exception:
            self._safe_terminate(sandbox)
            raise

        instance_id = str(uuid.uuid4())
        self._instances[instance_id] = {
            "sandbox": sandbox,
            "remote_work_dir": remote_work_dir,
            "language": language,
        }
        logger.info("Tenki sandbox created (instance=%s, session=%s)", instance_id, sandbox.id)

        return SandboxInstance(
            instance_id=instance_id,
            provider="tenki",
            status="running",
            metadata={"language": language, "remote_work_dir": remote_work_dir, "sandbox_id": sandbox.id},
        )

View on GitHub (pinned to 554fb1133a)

Solutions

1. Use the default image or ensure the custom image has a writable /home/tenki for user 'tenki'.
2. Increase disk_size_gb if the message indicates ENOSPC; check result.stderr_text in the exception message for the exact mkdir error.
3. Re-run create_instance() — transient microVM bootstrap failures are occasionally resolved by a fresh sandbox.

Example fix

# before
config = {"api_key": key, "image": "custom-no-tenki-user:1.0"}

# after
config = {"api_key": key}  # default image with writable /home/tenki

try:
    instance = provider.create_instance(template)
except RuntimeError as exc:
    if "Failed to create sandbox workspace" in str(exc):
        # sandbox was already terminated by the provider; safe to retry once
        instance = provider.create_instance(template)
    else:
        raise

When it happens

Trigger: A custom image where /home/tenki is missing or not writable by the unprivileged 'tenki' user, a read-only or full filesystem in the microVM, or the mkdir being killed by the min(timeout, 10)s deadline returning a failure exit code.

Common situations: Swapping the default image for one with a different user/home layout; disk_size_gb set so small the filesystem cannot hold the workspace; custom images without a 'tenki' user account.

Related errors

• Unsupported artifact entry: {path.name}
• Invalid Tenki provider configuration.
• Provider not initialized. Call initialize() first.
• Tenki quota exceeded: {exc}
• Tenki rate limited, please retry: {exc}

AI-assisted analysis of infiniflow/ragflow@554fb1133a (2026-08-15). Data as JSON: /api/errors/6cd0fccefe369751. Report an issue: GitHub.