wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-19 11:27:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix Serve long running test (#8223)

Browse Source
This commit is contained in:
Simon Mo
2020-04-29 14:09:33 -07:00
committed by SangBin Cho
parent c7c9db3840
commit 5d2e2532e7
1 changed files with 5 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ci/long_running_tests/workloads/serve.py
+5 -7
View File
@@ -36,7 +36,6 @@ ray.init(address=cluster.address, include_webui=True, webui_host="0.0.0.0")
serve.init(blocking=True, kv_store_connector=lambda ns: RayInternalKVStore(ns))
@serve.route("/echo")
@serve.accept_batch
def echo(_):
time.sleep(0.01) # Sleep for 10ms
@@ -44,15 +43,14 @@ def echo(_):
return ["hi {}".format(i) for i in range(serve.context.batch_size)]
print("Scaling to 30 replicas")
config = serve.get_backend_config("echo:v0")
config.num_replicas = 30
config.max_batch_size = 16
serve.set_backend_config("echo:v0", config)
serve.create_endpoint("echo", "/echo")
config = serve.BackendConfig(num_replicas=30, max_batch_size=16)
serve.create_backend(echo, "echo:v1", backend_config=config)
serve.set_traffic("echo", {"echo:v1": 1})
print("Warming up")
for _ in range(5):
resp = requests.get("http://127.0.0.1:8000/echo").json()
resp = requests.get("http://127.0.0.1:8000/echo").text
print(resp)
time.sleep(0.5)