# Copyright (C) 2003-2009 Robey Pointer # # This file is part of paramiko. # # Paramiko is free software; you can redistribute it and/or modify it under the # terms of the GNU Lesser General Public License as published by the Free # Software Foundation; either version 2.1 of the License, or (at your option) # any later version. # # Paramiko is distributed in the hope that it will be useful, but WITHOUT ANY # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR # A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more # details. # # You should have received a copy of the GNU Lesser General Public License # along with Paramiko; if not, write to the Free Software Foundation, Inc., # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. """ some unit tests to make sure sftp works well with large files. a real actual sftp server is contacted, and a new folder is created there to do test file operations in (so no existing files will be harmed). """ import random import struct import sys import time from paramiko.common import o660 from ._util import slow @slow class TestBigSFTP: def test_lots_of_files(self, sftp): """ create a bunch of files over the same session. """ numfiles = 100 try: for i in range(numfiles): target = f"{sftp.FOLDER}/file{i}.txt" with sftp.open(target, "w", 1) as f: f.write(f"this is file #{i}.\n") sftp.chmod(target, o660) # now make sure every file is there, by creating a list of filenmes # and reading them in random order. numlist = list(range(numfiles)) while len(numlist) > 0: r = numlist[random.randint(0, len(numlist) - 1)] with sftp.open(f"{sftp.FOLDER}/file{r}.txt") as f: assert f.readline() == f"this is file #{r}.\n" numlist.remove(r) finally: for i in range(numfiles): try: sftp.remove(f"{sftp.FOLDER}/file{i}.txt") except: pass def test_big_file(self, sftp): """ write a 1MB file with no buffering. """ kblob = 1024 * b"x" start = time.time() try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "w") as f: for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) end = time.time() sys.stderr.write(f"{round(end - start)}s") start = time.time() with sftp.open(f"{sftp.FOLDER}/hongry.txt", "r") as f: for n in range(1024): data = f.read(1024) assert data == kblob end = time.time() sys.stderr.write(f"{round(end - start)}s") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_big_file_pipelined(self, sftp): """ write a 1MB file, with no linefeeds, using pipelining. """ kblob = bytes().join([struct.pack(">H", n) for n in range(512)]) start = time.time() try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "wb") as f: f.set_pipelined(True) for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) end = time.time() sys.stderr.write(f"{round(end - start)}s") start = time.time() with sftp.open(f"{sftp.FOLDER}/hongry.txt", "rb") as f: file_size = f.stat().st_size f.prefetch(file_size) # read on odd boundaries to make sure the bytes aren't getting # scrambled n = 0 k2blob = kblob + kblob chunk = 629 size = 1024 * 1024 while n < size: if n + chunk > size: chunk = size - n data = f.read(chunk) offset = n % 1024 assert data == k2blob[offset : offset + chunk] n += chunk end = time.time() sys.stderr.write(f"{round(end - start)}s") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_prefetch_seek(self, sftp): kblob = bytes().join([struct.pack(">H", n) for n in range(512)]) try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "wb") as f: f.set_pipelined(True) for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) start = time.time() k2blob = kblob + kblob chunk = 793 for i in range(10): with sftp.open(f"{sftp.FOLDER}/hongry.txt", "rb") as f: file_size = f.stat().st_size f.prefetch(file_size) base_offset = (512 * 1024) + 17 * random.randint( 1000, 2000 ) offsets = [base_offset + j * chunk for j in range(100)] # randomly seek around and read them out for j in range(100): offset = offsets[random.randint(0, len(offsets) - 1)] offsets.remove(offset) f.seek(offset) data = f.read(chunk) n_offset = offset % 1024 assert data == k2blob[n_offset : n_offset + chunk] offset += chunk end = time.time() sys.stderr.write(f"{round(end - start)}s") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_readv_seek(self, sftp): kblob = bytes().join([struct.pack(">H", n) for n in range(512)]) try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "wb") as f: f.set_pipelined(True) for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) start = time.time() k2blob = kblob + kblob chunk = 793 for i in range(10): with sftp.open(f"{sftp.FOLDER}/hongry.txt", "rb") as f: base_offset = (512 * 1024) + 17 * random.randint( 1000, 2000 ) # make a bunch of offsets and put them in random order offsets = [base_offset + j * chunk for j in range(100)] readv_list = [] for j in range(100): o = offsets[random.randint(0, len(offsets) - 1)] offsets.remove(o) readv_list.append((o, chunk)) ret = f.readv(readv_list) for i in range(len(readv_list)): offset = readv_list[i][0] n_offset = offset % 1024 assert next(ret) == k2blob[n_offset : n_offset + chunk] end = time.time() sys.stderr.write(f"{round(end - start)}s") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_lots_of_prefetching(self, sftp): """ prefetch a 1MB file a bunch of times, discarding the file object without using it, to verify that paramiko doesn't get confused. """ kblob = 1024 * b"x" try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "w") as f: f.set_pipelined(True) for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) for i in range(10): with sftp.open(f"{sftp.FOLDER}/hongry.txt", "r") as f: file_size = f.stat().st_size f.prefetch(file_size) with sftp.open(f"{sftp.FOLDER}/hongry.txt", "r") as f: file_size = f.stat().st_size f.prefetch(file_size) for n in range(1024): data = f.read(1024) assert data == kblob if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_prefetch_readv(self, sftp): """ verify that prefetch and readv don't conflict with each other. """ kblob = bytes().join([struct.pack(">H", n) for n in range(512)]) try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "wb") as f: f.set_pipelined(True) for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) with sftp.open(f"{sftp.FOLDER}/hongry.txt", "rb") as f: file_size = f.stat().st_size f.prefetch(file_size) data = f.read(1024) assert data == kblob chunk_size = 793 base_offset = 512 * 1024 k2blob = kblob + kblob chunks = [ (base_offset + (chunk_size * i), chunk_size) for i in range(20) ] for data in f.readv(chunks): offset = base_offset % 1024 assert chunk_size == len(data) assert k2blob[offset : offset + chunk_size] == data base_offset += chunk_size sys.stderr.write(" ") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_large_readv(self, sftp): """ verify that a very large readv is broken up correctly and still returned as a single blob. """ kblob = bytes().join([struct.pack(">H", n) for n in range(512)]) try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "wb") as f: f.set_pipelined(True) for n in range(1024): f.write(kblob) if n % 128 == 0: sys.stderr.write(".") sys.stderr.write(" ") assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) with sftp.open(f"{sftp.FOLDER}/hongry.txt", "rb") as f: data = list(f.readv([(23 * 1024, 128 * 1024)])) assert len(data) == 1 data = data[0] assert len(data) == 128 * 1024 sys.stderr.write(" ") finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_big_file_big_buffer(self, sftp): """ write a 1MB file, with no linefeeds, and a big buffer. """ mblob = 1024 * 1024 * "x" try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "w", 128 * 1024) as f: f.write(mblob) assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") def test_big_file_renegotiate(self, sftp): """ write a 1MB file, forcing key renegotiation in the middle. """ t = sftp.sock.get_transport() t.packetizer.REKEY_BYTES = 512 * 1024 k32blob = 32 * 1024 * "x" try: with sftp.open(f"{sftp.FOLDER}/hongry.txt", "w", 128 * 1024) as f: for i in range(32): f.write(k32blob) assert ( sftp.stat(f"{sftp.FOLDER}/hongry.txt").st_size == 1024 * 1024 ) assert t.H != t.session_id # try to read it too. with sftp.open(f"{sftp.FOLDER}/hongry.txt", "r", 128 * 1024) as f: file_size = f.stat().st_size f.prefetch(file_size) total = 0 while total < 1024 * 1024: total += len(f.read(32 * 1024)) finally: sftp.remove(f"{sftp.FOLDER}/hongry.txt") t.packetizer.REKEY_BYTES = pow(2, 30)