TapTap/FastSync
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge branch 'main' into metadata-transfer, resolve conflicts

Browse Source 
Conflicts resolved:
- config_create signature: added both use_metadata and use_sendfile params
- config_send/config_receive: wire protocol includes both use_metadata and use_sendfile
- client.c: config_create call updated, arg parsing includes both -M/--preserve and -f/--sendfile
- test_config.c: all three config_create calls updated
- file_send_sendfile: use file->data->size instead of removed struct stat
This commit is contained in:
TapTap
2026-07-05 20:29:04 +02:00
parent d4fdc16c9e 1f75522d21
commit 660834d453
9 changed files with 383 additions and 65 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test.py
+308 -48
View File
@@ -1,11 +1,13 @@
import argparse
import filecmp
import os
import re
import shutil
import subprocess
import sys
import tempfile
import time
import socket
TEST_DIR = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
DEFAULT_SOURCE_DIR = os.path.join(TEST_DIR, "source")
@@ -17,14 +19,23 @@ base_client_cmd = ["./build/client"]
DISK_DEVICE = "/dev/nvme0n1p5"
READ_BPS_MAX = "15M"
WRITE_BPS_MAX = "10M"
NET_LIMIT = "100mbit"
NET_DELAY = "100ms"
NETWORK_INTERFACE = "lo"
NET_LIMIT_CMD = (
f"sudo tc qdisc add dev {NETWORK_INTERFACE} root netem rate {NET_LIMIT} delay {NET_DELAY}".split()
)
NET_RESET_CMD = f"sudo tc qdisc del dev {NETWORK_INTERFACE} root".split()
NETWORK_PROFILES = {
"Unlimited": {},
"LAN": {
"rate": "1000mbit",
"delay": "20ms",
"jitter": "1ms",
"loss": "0.1%",
},
"WAN": {
"rate": "100mbit",
"delay": "50ms",
"jitter": "10ms",
"loss": "1%",
},
}
CLIENT_CMD_PREFIX = [
"sudo",
@@ -37,7 +48,7 @@ CLIENT_CMD_PREFIX = [
]
TEST_CASES = [
{"name": "Standard (Single-threaded)", "flags": []},
{"name": "Standard", "flags": []},
{"name": "Multithreading (-m)", "flags": ["-m"]},
{"name": "Compression (-c)", "flags": ["-c"]},
{"name": "Chunk Serialization (-s)", "flags": ["-s"]},
@@ -48,8 +59,41 @@ TEST_CASES = [
"name": "Multithreading + Compression + Chunk Serialization (-m -c -s)",
"flags": ["-m", "-c", "-s"],
},
{"name": "Sendfile (-f)", "flags": ["-f"]},
{"name": "Sendfile + Multithreading (-f -m)", "flags": ["-f", "-m"]},
]
RSYNC_CASES = [
{"name": "rsync (archive)", "args": ["-aH"]},
{"name": "rsync (archive + compress)", "args": ["-aHz"]},
]
def netem_apply(profile):
params = NETWORK_PROFILES[profile]
if not params:
netem_reset()
return
netem_reset()
cmd = ["sudo", "tc", "qdisc", "add", "dev", NETWORK_INTERFACE, "root", "netem"]
cmd += ["rate", params["rate"]]
cmd += ["delay", params["delay"], params["jitter"]]
cmd += ["loss", params["loss"]]
subprocess.run(cmd, check=True, capture_output=True)
def netem_reset():
subprocess.run(
f"sudo tc qdisc del dev {NETWORK_INTERFACE} root".split(),
capture_output=True,
)
def find_free_port():
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
s.bind(('', 0))
return s.getsockname()[1]
def generate_test_files(source_dir):
if os.path.exists(source_dir):
@@ -86,14 +130,14 @@ def generate_test_files(source_dir):
total_mb = written / (1024 * 1024)
print(f" Generated {total_mb:.1f} MB of test data in {source_dir}")
return written
def verify_transfer(source_dir, dest_dir):
def verify_transfer(source_dir, received_dir):
source_dir = os.path.abspath(source_dir)
dest_dir = os.path.abspath(dest_dir)
received_dir = os.path.abspath(received_dir)
received_prefix = os.path.join(dest_dir, source_dir.lstrip(os.sep))
if not os.path.exists(received_prefix):
if not os.path.exists(received_dir):
return [], ["no received files found"]
mismatches = []
@@ -103,7 +147,7 @@ def verify_transfer(source_dir, dest_dir):
for f in files:
src_path = os.path.join(root, f)
rel = os.path.relpath(src_path, source_dir)
dst_path = os.path.join(received_prefix, rel)
dst_path = os.path.join(received_dir, rel)
if not os.path.exists(dst_path):
missing.append(rel)
@@ -113,25 +157,28 @@ def verify_transfer(source_dir, dest_dir):
return mismatches, missing
def run_suite(env_name, apply_limits, source_dir, dest_dir):
def run_profile(profile_name, source_dir, dest_dir):
results = []
is_limited = profile_name != "Unlimited"
params = NETWORK_PROFILES[profile_name]
print(f"\n{'=' * 60}")
print(f"Suite: {env_name}")
print(f"Profile: {profile_name}")
print(f"{'=' * 60}")
if apply_limits:
if is_limited:
p = params
print(f" Network: rate={p['rate']}, delay={p['delay']} ±{p['jitter']}, loss={p['loss']}")
print(f" Disk I/O: Reads <= {READ_BPS_MAX}, Writes <= {WRITE_BPS_MAX}")
print(f" Network: {NET_LIMIT}, {NET_DELAY} delay")
client_prefix = CLIENT_CMD_PREFIX
else:
print(" Baseline (no limits)")
print(" No limits applied")
client_prefix = []
try:
if apply_limits:
subprocess.run(NET_LIMIT_CMD, check=True)
if is_limited:
netem_apply(profile_name)
else:
subprocess.run(NET_RESET_CMD, capture_output=True)
netem_reset()
for case in TEST_CASES:
name = case["name"]
@@ -175,11 +222,12 @@ def run_suite(env_name, apply_limits, source_dir, dest_dir):
mismatches, missing = [], []
if client_result.returncode == 0:
mismatches, missing = verify_transfer(source_dir, dest_dir)
received = os.path.join(dest_dir, os.path.abspath(source_dir).lstrip(os.sep))
mismatches, missing = verify_transfer(source_dir, received)
entry = {
"name": name,
"suite": env_name,
"suite": profile_name,
"time": f"{duration:.4f}s" if client_result.returncode == 0 else "N/A",
}
@@ -210,11 +258,11 @@ def run_suite(env_name, apply_limits, source_dir, dest_dir):
except subprocess.TimeoutExpired:
results.append(
{"name": name, "suite": env_name, "status": "Timeout", "time": "N/A", "error": "Exceeded 15s"}
{"name": name, "suite": profile_name, "status": "Timeout", "time": "N/A", "error": "Exceeded 15s"}
)
except Exception as e:
results.append(
{"name": name, "suite": env_name, "status": "Error", "time": "N/A", "error": str(e)}
{"name": name, "suite": profile_name, "status": "Error", "time": "N/A", "error": str(e)}
)
finally:
if server_process:
@@ -224,18 +272,175 @@ def run_suite(env_name, apply_limits, source_dir, dest_dir):
server_process.kill()
server_process.wait()
except subprocess.CalledProcessError as e:
print(f" Error running limit command: {' '.join(e.cmd)}")
finally:
if apply_limits:
# Rsync tests (over network via daemon, so tc netem applies)
rsync_port = find_free_port()
rsyncd_conf = os.path.join(tempfile.gettempdir(), f"rsyncd-{rsync_port}.conf")
with open(rsyncd_conf, "w") as f:
f.write(f"""port = {rsync_port}
read only = yes
[source]
path = {source_dir}
""")
rsync_daemon = None
try:
rsync_daemon = subprocess.Popen(
["rsync", "--daemon", "--no-detach", f"--config={rsyncd_conf}"],
stdout=subprocess.DEVNULL, stderr=None
)
for _ in range(50):
time.sleep(0.1)
if rsync_daemon.poll() is not None:
print(f" rsync daemon exited (rc={rsync_daemon.returncode})")
raise RuntimeError("rsync daemon failed to start")
try:
with socket.create_connection(("127.0.0.1", rsync_port), timeout=0.3):
break
except (ConnectionRefusedError, OSError):
continue
else:
print(f" timed out waiting for rsync daemon on port {rsync_port}")
raise RuntimeError("rsync daemon did not start")
for case in RSYNC_CASES:
name = case["name"]
rsync_args = case["args"]
print(f"\n --- {name} ---")
if os.path.exists(dest_dir):
shutil.rmtree(dest_dir)
try:
rsync_cmd = (
client_prefix
+ ["rsync"]
+ rsync_args
+ [f"rsync://localhost:{rsync_port}/source/", f"{dest_dir}/"]
)
print(f" Running: {' '.join(rsync_cmd)}")
start_time = time.monotonic()
rsync_result = subprocess.run(
rsync_cmd, capture_output=True, text=True, timeout=120
)
end_time = time.monotonic()
duration = end_time - start_time
mismatches, missing = [], []
if rsync_result.returncode == 0:
mismatches, missing = verify_transfer(source_dir, dest_dir)
entry = {
"name": name,
"suite": profile_name,
"time": f"{duration:.4f}s" if rsync_result.returncode == 0 else "N/A",
}
if rsync_result.returncode == 0 and not mismatches and not missing:
entry["status"] = "Success"
entry["error"] = ""
else:
entry["status"] = "Failed"
errors = []
if rsync_result.returncode != 0:
errs = []
for line in (rsync_result.stderr or "").split("\n"):
line = line.strip()
if not line:
continue
if line.startswith("Running as unit:"):
continue
errs.append(line)
for line in (rsync_result.stdout or "").split("\n"):
line = line.strip()
if not line:
continue
errs.append(line)
if not errs:
errs.append("No output")
err = " | ".join(errs[-3:])
errors.append(f"Exit code {rsync_result.returncode}: {err[:200]}")
# Retry without systemd-run to reveal the actual error
if client_prefix:
tmp_dest = tempfile.mkdtemp()
try:
plain = subprocess.run(
["rsync", "-aH", f"rsync://localhost:{rsync_port}/source/", f"{tmp_dest}/"],
capture_output=True, text=True, timeout=30
)
if plain.returncode != 0:
plain_errs = [l for l in (plain.stderr or "").split("\n") if l.strip()]
if plain_errs:
errors.append(f"raw: {plain_errs[-1][:150]}")
finally:
shutil.rmtree(tmp_dest, ignore_errors=True)
if missing:
errors.append(f"Missing ({len(missing)}): {', '.join(missing[:5])}")
if mismatches:
errors.append(f"Mismatch ({len(mismatches)}): {', '.join(mismatches[:3])}")
entry["error"] = " | ".join(errors)
results.append(entry)
except subprocess.TimeoutExpired:
results.append(
{"name": name, "suite": profile_name, "status": "Timeout", "time": "N/A", "error": "Exceeded 120s"}
)
except Exception as e:
results.append(
{"name": name, "suite": profile_name, "status": "Error", "time": "N/A", "error": str(e)}
)
finally:
if rsync_daemon:
try:
rsync_daemon.wait(timeout=5)
except subprocess.TimeoutExpired:
rsync_daemon.kill()
rsync_daemon.wait()
try:
subprocess.run(NET_RESET_CMD, check=True, capture_output=True)
os.unlink(rsyncd_conf)
except Exception:
pass
except subprocess.CalledProcessError as e:
print(f" Error running netem command: {' '.join(e.cmd)}")
finally:
if is_limited:
try:
netem_reset()
except Exception:
pass
return results
def parse_rate_to_bytes_per_sec(rate_str):
m = re.match(r'(\d+)\s*(mbit|gbit|kbit|bit)', rate_str)
if not m:
return None
val = int(m.group(1))
unit = m.group(2)
bits_per_sec = {
'bit': val,
'kbit': val * 1000,
'mbit': val * 1_000_000,
'gbit': val * 1_000_000_000,
}.get(unit)
return bits_per_sec / 8 if bits_per_sec is not None else None
def format_throughput(bps):
if bps >= 1_000_000_000:
return f"{bps/1_000_000_000:.1f} GB/s"
if bps >= 1_000_000:
return f"{bps/1_000_000:.1f} MB/s"
if bps >= 1_000:
return f"{bps/1_000:.1f} KB/s"
return f"{bps:.0f} B/s"
def main():
parser = argparse.ArgumentParser(description="FastSync integration test / benchmark")
parser.add_argument("--source-dir", default=DEFAULT_SOURCE_DIR,
@@ -244,10 +449,10 @@ def main():
help="Destination directory for received files (default: %(default)s)")
parser.add_argument("--keep-data", action="store_true",
help="Keep test_data directory after run")
parser.add_argument("--no-throttled", action="store_true",
help="Skip throttled suite (requires sudo)")
parser.add_argument("--no-unlimited", action="store_true",
help="Skip unlimited suite")
parser.add_argument("--unlimited", action="store_true",
help="Run Unlimited profile instead of LAN (no network limits)")
parser.add_argument("--wan", action="store_true",
help="Run WAN profile instead of LAN (100mbit, 50ms, 1% loss)")
args = parser.parse_args()
os.system("cmake -B build -S . > /dev/null 2>&1")
@@ -256,33 +461,88 @@ def main():
print("Build failed")
sys.exit(1)
generate_test_files(args.source_dir)
total_bytes = generate_test_files(args.source_dir)
if os.path.exists(args.dest_dir):
shutil.rmtree(args.dest_dir)
os.makedirs(args.dest_dir, exist_ok=True)
profiles_to_run = []
if args.unlimited:
profiles_to_run.append("Unlimited")
elif args.wan:
profiles_to_run.append("WAN")
else:
profiles_to_run.append("LAN")
try:
all_results = []
if not args.no_unlimited:
for profile in profiles_to_run:
all_results.extend(
run_suite("Unlimited", False, args.source_dir, args.dest_dir)
run_profile(profile, args.source_dir, args.dest_dir)
)
if not args.no_throttled:
all_results.extend(
run_suite("Throttled", True, args.source_dir, args.dest_dir)
)
print("\n" + "=" * 110)
print(f"{'RESULTS':^110}")
print("=" * 110)
print(f"{'Configuration':<45} | {'Suite':<12} | {'Status':<8} | {'Time':<10} | {'Details'}")
print("-" * 110)
print("\n" + "=" * 130)
print(f"{'RESULTS':^130}")
print("=" * 130)
print(f"{'Configuration':<45} | {'Profile':<12} | {'Status':<8} | {'Time':<10} | {'Details'}")
print("-" * 130)
for res in all_results:
print(
f"{res['name']:<45} | {res['suite']:<12} | {res['status']:<8} | {res['time']:<10} | {res['error']}"
)
# --- Additional metrics per profile ---
profiles_results = {}
for res in all_results:
profiles_results.setdefault(res["suite"], []).append(res)
for profile_name, results in profiles_results.items():
params = NETWORK_PROFILES.get(profile_name)
if not params or "rate" not in params:
continue
client_times = []
rsync_times = {}
for r in results:
if r["status"] != "Success" or r["time"] == "N/A":
continue
t = float(r["time"].rstrip("s"))
if r["name"].startswith("rsync"):
rsync_times[r["name"]] = t
else:
client_times.append((t, r["name"]))
if not client_times or len(rsync_times) < 2:
continue
best_time, best_name = min(client_times, key=lambda x: x[0])
rate_Bps = parse_rate_to_bytes_per_sec(params["rate"])
theoretical_max_time = None
speedup_vs_theoretical = None
if rate_Bps is not None:
theoretical_max_time = total_bytes / rate_Bps
speedup_vs_theoretical = theoretical_max_time / best_time
print(f"\n {'─' * 90}")
print(f" Profile: {profile_name}")
print(f" {'─' * 90}")
print(f" Total data size: {total_bytes / (1024*1024):.1f} MB")
print(f" Network rate: {params['rate']} ({format_throughput(rate_Bps)})" if rate_Bps else "")
print(f" Best client configuration: {best_name}")
print(f" Best client time: {best_time:.4f}s")
if theoretical_max_time is not None:
print(f" Theoretical max (uncompressed): {theoretical_max_time:.4f}s")
print(f" Speedup vs theoretical max: {speedup_vs_theoretical:.2f}x")
rsync_archive = rsync_times.get("rsync (archive)")
rsync_compress = rsync_times.get("rsync (archive + compress)")
if rsync_archive:
print(f" Speedup vs rsync (archive): {rsync_archive / best_time:.2f}x")
if rsync_compress:
print(f" Speedup vs rsync (compress): {rsync_compress / best_time:.2f}x")
failed = [r for r in all_results if r["status"] != "Success"]
if failed:
print(f"\n {len(failed)} test(s) FAILED")