scorpiofs 0.2.2

FUSE-based virtual filesystem with Antares overlay for monorepo builds
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200

# Performance test for FUSE.
#!/usr/bin/env python3
import os
import time
import random
import argparse
import csv
import subprocess
from multiprocessing import Pool, cpu_count
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np

class FusePerfTester:
    def __init__(self, test_path, file_size='1M', output='report'):
        self.test_path = test_path
        self.file_size = self.parse_size(file_size)
        self.output_dir = output
        self.stats = {
            'write': {'time': [], 'speed': []},
            'read': {'time': [], 'speed': []},
            'random_read': {'iops': []},
            'metadata': {'create': [], 'delete': [], 'stat': []}
        }
        os.makedirs(self.output_dir, exist_ok=True)

    @staticmethod
    def parse_size(size_str):
        units = {"B": 1, "K": 2**10, "M": 2**20, "G": 2**30}
        num = int(''.join(filter(str.isdigit, size_str)))
        unit = ''.join(filter(str.isalpha, size_str.upper())) or 'B'
        return num * units[unit]

    def _cleanup(self):
        subprocess.run(['rm', '-rf', os.path.join(self.test_path, 'testfile*')], 
                      check=True)

    def sequential_write(self, iterations=10):
        for i in range(iterations):
            filename = os.path.join(self.test_path, f'testfile_write_{i}.dat')
            start = time.time()
            with open(filename, 'wb') as f:
                f.write(os.urandom(self.file_size))
            elapsed = time.time() - start
            speed = self.file_size / elapsed / (1024**2)  # MB/s
            self.stats['write']['time'].append(elapsed)
            self.stats['write']['speed'].append(speed)

    def sequential_read(self, iterations=10):
        # 
        test_file = os.path.join(self.test_path, 'testfile_read.dat')
        with open(test_file, 'wb') as f:
            f.write(os.urandom(self.file_size))

        for _ in range(iterations):
            start = time.time()
            with open(test_file, 'rb') as f:
                f.read()
            elapsed = time.time() - start
            speed = self.file_size / elapsed / (1024**2)
            self.stats['read']['time'].append(elapsed)
            self.stats['read']['speed'].append(speed)

    def random_io_test(self, block_size=4096, duration=30):
        """randoem IOPS test"""
        test_file = os.path.join(self.test_path, 'testfile_random.dat')
        file_size = 1 * 1024**3  # 1GB for random test
        with open(test_file, 'wb') as f:
            f.truncate(file_size)

        blocks = file_size // block_size
        count = 0
        start = time.time()
        while time.time() - start < duration:
            with open(test_file, 'rb+') as f:
                offset = random.randint(0, blocks-1) * block_size
                f.seek(offset)
                f.write(os.urandom(block_size))
            count += 1
        self.stats['random_read']['iops'].append(count // duration)

    def metadata_ops(self, iterations=1000):
        """meta data operation test"""
        for i in range(iterations//3):
            # File create
            start = time.time()
            with open(os.path.join(self.test_path, f'temp_{i}.dat'), 'w') as f:
                f.write('test')
            self.stats['metadata']['create'].append(time.time() - start)

            # stat 
            start = time.time()
            os.stat(os.path.join(self.test_path, f'temp_{i}.dat'))
            self.stats['metadata']['stat'].append(time.time() - start)

            # delete 
            start = time.time()
            os.remove(os.path.join(self.test_path, f'temp_{i}.dat'))
            self.stats['metadata']['delete'].append(time.time() - start)

    def concurrent_test(self, workers=cpu_count()):
        """multi-thread test"""
        test_file = os.path.join(self.test_path, 'concurrent.dat')
        file_size = 100 * 1024**2  # 100MB
        with open(test_file, 'wb') as f:
            f.truncate(file_size)

        def worker(_):
            block_size = 4096
            offset = random.randint(0, (file_size//block_size)-1) * block_size
            data = os.urandom(block_size)
            with open(test_file, 'r+b') as f:
                f.seek(offset)
                f.write(data)

        start = time.time()
        with Pool(workers) as p:
            p.map(worker, range(1000))
        elapsed = time.time() - start
        self.stats['concurrency'] = {'total_time': elapsed, 'workers': workers}

    def generate_report(self):
        """view the result"""
        sns.set(style="whitegrid")

        # throughput test
        plt.figure(figsize=(12, 6))
        write_speeds = self.stats['write']['speed']
        read_speeds = self.stats['read']['speed']
        sns.lineplot(x=range(len(write_speeds)), y=write_speeds, label='Write Speed (MB/s)')
        sns.lineplot(x=range(len(read_speeds)), y=read_speeds, label='Read Speed (MB/s)')
        plt.title('Sequential R/W Throughput')
        plt.savefig(os.path.join(self.output_dir, 'throughput.png'))
        plt.close()

        # metadata delay
        plt.figure(figsize=(10,5))
        data = [
            np.mean(self.stats['metadata']['create']) * 1000,
            np.mean(self.stats['metadata']['stat']) * 1000,
            np.mean(self.stats['metadata']['delete']) * 1000
        ]
        labels = ['Create', 'Stat', 'Delete']
        sns.barplot(x=labels, y=data)
        plt.ylabel('Latency (ms)')
        plt.title('Metadata Operation Latency')
        plt.savefig(os.path.join(self.output_dir, 'metadata.png'))
        plt.close()

        # save key data to CSV
        with open(os.path.join(self.output_dir, 'summary.csv'), 'w') as f:
            writer = csv.writer(f)
            writer.writerow(['Test Item', 'Average', 'Max', 'Min', 'Std'])
            for test in ['write', 'read']:
                data = self.stats[test]['speed']
                writer.writerow([
                    f"{test.capitalize()} Speed (MB/s)",
                    np.mean(data),
                    max(data),
                    min(data),
                    np.std(data)
                ])
            if 'concurrency' in self.stats:
                writer.writerow([
                    'Concurrency (ops/s)', 
                    1000 / self.stats['concurrency']['total_time'],
                    '', '', ''
                ])

        # clean
        self._cleanup()

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='FUSE Filesystem Performance Tester')
    parser.add_argument('path', help='FUSE mount path to test')
    parser.add_argument('-s', '--size', default='128M', 
                      help='Test file size (e.g. 1M, 512K)')
    parser.add_argument('-o', '--output', default='report', 
                      help='Output directory for reports')
    args = parser.parse_args()

    tester = FusePerfTester(args.path, args.size, args.output)
    
    print("Running sequential write test...")
    tester.sequential_write()
    
    print("Running sequential read test...")
    tester.sequential_read()
    
    print("Running random I/O test...")
    tester.random_io_test()
    
    print("Testing metadata operations...")
    tester.metadata_ops()
    
    print("Testing concurrency...")
    tester.concurrent_test()
    
    print("Generating report...")
    tester.generate_report()