Refactor GPU environment verification and SSH execution for improved flexibility
- Removed the direct injection of environment variables into the script, preventing issues with `from __future__` imports. - Updated the `run_python` function to accept an `env` parameter for passing runtime environment variables, enhancing script execution control. - Adjusted tests to verify the new behavior, ensuring that environment variables are correctly set without modifying the script content.
This commit is contained in:
@@ -323,7 +323,6 @@ def verify_gpu_env(
|
||||
encoding="utf-8"
|
||||
)
|
||||
swarm_flag = "1" if want_swarm else "0"
|
||||
prefix = f"import os\nos.environ['GPU_RENT_CHECK_SWARM']={swarm_flag!r}\n"
|
||||
log(
|
||||
"проверка GPU-стека: nvidia-smi, CUDA"
|
||||
+ (", torch в Comfy venv" if want_swarm else " (llm-only — без torch)")
|
||||
@@ -337,10 +336,11 @@ def verify_gpu_env(
|
||||
out = run_python(
|
||||
cfg,
|
||||
host,
|
||||
prefix + script,
|
||||
script,
|
||||
remote_path="/tmp/gpu-rent-stack_env_probe.py",
|
||||
timeout=180,
|
||||
log=None,
|
||||
env={"GPU_RENT_CHECK_SWARM": swarm_flag},
|
||||
)
|
||||
except Exception as exc:
|
||||
last = [ServiceCheck("gpu-env", False, str(exc)[:200], "vm")]
|
||||
|
||||
+16
-2
@@ -352,8 +352,13 @@ def run_python(
|
||||
remote_path: str,
|
||||
timeout: int = 1800,
|
||||
log: Callable[[str], None] | None = None,
|
||||
env: dict[str, str] | None = None,
|
||||
) -> str:
|
||||
"""Upload a Python script and run it as SSH user (not root)."""
|
||||
"""Upload a Python script and run it as SSH user (not root).
|
||||
|
||||
Do not prepend Python code to ``script`` (breaks ``from __future__``).
|
||||
Pass runtime flags via ``env`` instead.
|
||||
"""
|
||||
client = _connect(cfg, host)
|
||||
chunks: list[str] = []
|
||||
try:
|
||||
@@ -362,7 +367,16 @@ def run_python(
|
||||
fh.write(script)
|
||||
sftp.chmod(remote_path, 0o755)
|
||||
sftp.close()
|
||||
command = f"python3 {shlex.quote(remote_path)}"
|
||||
exports = ""
|
||||
if env:
|
||||
parts = [
|
||||
f"{key}={shlex.quote(str(val))}"
|
||||
for key, val in env.items()
|
||||
if key
|
||||
]
|
||||
if parts:
|
||||
exports = " ".join(parts) + " "
|
||||
command = f"{exports}python3 {shlex.quote(remote_path)}"
|
||||
_stdin, stdout, stderr = client.exec_command(command, timeout=timeout, get_pty=True)
|
||||
_stream_pty_output(stdout, log=log, chunks=chunks)
|
||||
code = stdout.channel.recv_exit_status()
|
||||
|
||||
Reference in New Issue
Block a user