Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,7 @@ main
.devcontainer
**/*.tgz
**/.DS_Store
.claude/


# python
Expand Down
20 changes: 14 additions & 6 deletions python/patio/topo/client/sgl_topo_client.py
Original file line number Diff line number Diff line change
Expand Up @@ -59,8 +59,6 @@ def get_health_check_endpoint(worker_info: dict) -> Optional[str]:

return f"{local_url}:{port}"

def translate_http_response_parameters(raw: str) -> str:
return raw.replace("/", "%2F")

class SGLangGroupTopoClient(GroupTopoClient):
_instance = None
Expand All @@ -77,6 +75,7 @@ def __init__(self, worker_info: dict):
self.health_check_endpoint = get_health_check_endpoint(worker_info)
self.worker_endpoint = get_worker_endpoint(worker_info)
self.sgl_router_endpoint = get_sgl_router_endpoint(worker_info)
self.worker_id = None
self.__initialized = True

def wait_engine_ready(self, worker_info: dict) -> bool:
Expand Down Expand Up @@ -117,7 +116,13 @@ def f():
resp = requests.post(worker_registration_url, json=worker_info, headers={"Content-Type": "application/json"})
if resp.status_code == 202:
# Status Code 202 Accepted
logger.info(f"registered worker successfully.")
self.worker_id = resp.json().get("worker_id")
if self.worker_id is None:
raise Exception(
f"register failed: missing worker_id in response body, "
f"url: {worker_registration_url}, status_code: {resp.status_code}, content: {resp.text}"
)
logger.info(f"registered worker successfully. worker_id: {self.worker_id}")
else:
raise Exception(f"register failed, url: {worker_registration_url}, status_code: {resp.status_code}, content: {resp.text}")

Expand All @@ -131,13 +136,16 @@ def f():


def unregister(self):
if self.worker_id is None:
logger.warning("worker_id is not set, skipping unregister (registration may have failed)")
return False

def f():
url = f"http://{self.worker_endpoint}"
worker_registration_url = f"http://{self.sgl_router_endpoint}/workers/{translate_http_response_parameters(url)}"
worker_registration_url = f"http://{self.sgl_router_endpoint}/workers/{self.worker_id}"
resp = requests.delete(worker_registration_url)
if resp.status_code == 202:
# Status Code 202 Accepted
logger.info(f"unregistered worker successfully.")
logger.info(f"unregistered worker successfully. worker_id: {self.worker_id}")
else:
raise Exception(
f"unregister failed, url: {worker_registration_url}, status_code: {resp.status_code}, content: {resp.text}")
Expand Down
Loading