Files
qdrant/tests/consensus_tests/test_triple_replication.py
Andrey Vasnetsov 926a058c73 Fixing missed leader inconsistency (#1286)
* WIP: introduce local state

* WIP: sync local state with consensus in idle

* fmt

* rm extensive debug

* update triple replication test

* update triple replication test

* rm debug logs

* fix check for established leader + only sync local if no proposals

* remove unused file

* test compatible with python 3.8

* test compatible with python 3.8

* longer wait for consensus

* longer wait for consensus

* extra sleep in test

* remove extra sleep

* Fixing missed leader inconsistency - transfers (#1298)

* explicit request timeout

* explicit request timeout

* explicit request timeout

* explicit request timeout

* explicit request timeout

* prevent double handelling of the transfer termination

* kill the process

* log on inconsistency

* log on inconsistency

* log on inconsistency

* debug

* revert debug in test

* forward updates to partial shards, abort transactions on dead node report

* fmt

* disable retry of transfer, if the transfer was cancelled + allow predictable ports in test

* fix import
2022-12-23 00:45:04 +01:00

98 lines
3.1 KiB
Python

import multiprocessing
import pathlib
from .fixtures import upsert_random_points, create_collection
from .utils import *
from .assertions import assert_http_ok
N_PEERS = 3
N_SHARDS = 1
N_REPLICA = 3
def check_collection_cluster(peer_url, collection_name):
res = requests.get(f"{peer_url}/collections/{collection_name}/cluster", timeout=10)
assert_http_ok(res)
return res.json()["result"]['local_shards'][0]
def update_points_in_loop(peer_url, collection_name):
offset = 0
limit = 3
while True:
upsert_random_points(peer_url, limit, collection_name, offset=offset)
offset += limit
# time.sleep(0.01)
def run_update_points_in_background(peer_url, collection_name):
p = multiprocessing.Process(target=update_points_in_loop, args=(peer_url, collection_name))
p.start()
return p
def test_triple_replication(tmp_path: pathlib.Path):
assert_project_root()
peer_api_uris, peer_dirs, bootstrap_uri = start_cluster(tmp_path, N_PEERS)
create_collection(peer_api_uris[0], shard_number=N_SHARDS, replication_factor=N_REPLICA)
wait_collection_exists_and_active_on_all_peers(
collection_name="test_collection",
peer_api_uris=peer_api_uris
)
upload_process = run_update_points_in_background(peer_api_uris[1], "test_collection")
time.sleep(0.3)
# Kill leader peer
killed_id = 0
p = processes.pop(killed_id)
p.kill()
peer_api_uris.pop(killed_id)
new_url = start_peer(peer_dirs[killed_id], f"peer_{killed_id}_restarted.log", bootstrap_uri)
peer_api_uris.append(new_url)
wait_for_peer_online(new_url)
time.sleep(0.3)
upload_process.kill()
timeout = 10
while True:
if timeout < 0:
raise Exception("Timeout waiting for all replicas to be active")
all_active = True
points_counts = set()
for peer_api_uri in peer_api_uris:
res = check_collection_cluster(peer_api_uri, "test_collection")
points_counts.add(res['points_count'])
if res['state'] != 'Active':
all_active = False
if all_active:
if len(points_counts) != 1:
with open("test_triple_replication.log", "w") as f:
for peer_api_uri in peer_api_uris:
collection_name = "test_collection"
res = requests.get(f"{peer_api_uri}/collections/{collection_name}/cluster", timeout=10)
f.write(f"{peer_api_uri} {res.json()['result']}\n")
for peer_api_uri in peer_api_uris:
res = requests.get(f"{peer_api_uri}/cluster", timeout=10)
f.write(f"{peer_api_uri} {res.json()['result']}\n")
for peer_api_uri in peer_api_uris:
res = requests.post(f"{peer_api_uri}/collections/test_collection/points/count", json={"exact": True})
print(res.json())
assert False, f"Points count is not equal on all peers: {points_counts}"
break
time.sleep(1)
timeout -= 1