并发追加到文件：写入丢失

Question

并发追加到文件支持功能吗？

我使用并发线程 + 每个线程的 fstream 对此进行了测试。我看到数据没有损坏，但是一些写入丢失了。 写入完成后文件大小小于预期。写入不会重叠。

如果我在每个 fstream 中使用自定义搜索来编写每个线程将写入的偏移量，则不会丢失任何写入。

这里是示例代码：

#include <fstream>
#include <vector>
#include <thread>
#include "gtest/gtest.h"

void append_concurrently(string filename, const int data_in_gb, const int num_threads, const char start_char,
    bool stream_cache = true) {
    const int offset = 1024;
    const long long num_records_each_thread = (data_in_gb * 1024 * ((1024 * 1024) / (num_threads * offset)));

    {
        auto write_file_fn = [&](int index) {
            // each thread has its own handle
            fstream file_handle(filename, fstream::app | fstream::binary);
            if (!stream_cache) {
                file_handle.rdbuf()->pubsetbuf(nullptr, 0); // no bufferring in fstream
            }

            vector<char> data(offset, (char)(index + start_char));

            for (long long i = 0; i < num_records_each_thread; ++i) {
                file_handle.write(data.data(), offset);
                if (!file_handle) {
                    std::cout << "File write failed: "
                        << file_handle.fail() << " " << file_handle.bad() << " " << file_handle.eof() << std::endl;
                    break;
                }
            }

            // file_handle.flush();
        };

        auto start_time = chrono::high_resolution_clock::now();
        vector<thread> writer_threads;
        for (int i = 0; i < num_threads; ++i) {
            writer_threads.push_back(std::thread(write_file_fn, i));
        }

        for (int i = 0; i < num_threads; ++i) {
            writer_threads[i].join();
        }

        auto end_time = chrono::high_resolution_clock::now();

        std::cout << filename << " Data written : " << data_in_gb << " GB, " << num_threads << " threads "
            << ", cache " << (stream_cache ? "true " : "false ") << ", size " << offset << " bytes ";
        std::cout << "Time taken: " << (end_time - start_time).count() / 1000 << " micro-secs" << std::endl;
    }

    {
        ifstream file(filename, fstream::in | fstream::binary);
        file.seekg(0, ios_base::end);
        
        // This EXPECT_EQ FAILS as file size is smaller than EXPECTED
        EXPECT_EQ(num_records_each_thread * num_threads * offset, file.tellg());
        file.seekg(0, ios_base::beg);
        EXPECT_TRUE(file);

        char data[offset]{ 0 };
        for (long long i = 0; i < (num_records_each_thread * num_threads); ++i) {
            file.read(data, offset);
            EXPECT_TRUE(file || file.eof()); // should be able to read until eof
            char expected_char = data[0]; // should not have any interleaving of data.

            bool same = true;
            for (auto & c : data) {
                same = same && (c == expected_char) && (c != 0);
            }

            EXPECT_TRUE(same); // THIS PASSES
            if (!same) {
                std::cout << "corruption detected !!!" << std::endl;
                break;
            }

            if (file.eof()) { // THIS FAILS as file size is smaller
                EXPECT_EQ(num_records_each_thread * num_threads, i + 1);
                break;
            }
        }
    }
}

TEST(fstream, file_concurrent_appends) {
    string filename = "file6.log";
    const int data_in_gb = 1;
    
    {
        // trunc file before write threads start.
        {
            fstream file(filename, fstream::in | fstream::out | fstream::trunc | fstream::binary);
        }
        append_concurrently(filename, data_in_gb, 4, 'B', false);
    }
    std::remove(filename.c_str());
}

编辑：

我将 fstream 移动到由所有线程共享。现在，对于 512 字节的缓冲区大小，我看到 8 次写入，总共丢失了 4 KB。

    const int offset = 512;
    const long long num_records_each_thread = (data_in_gb * 1024 * ((1024 * 1024) / (num_threads * offset)));

    fstream file_handle(filename, fstream::app | fstream::binary);
    if (!stream_cache) {
        file_handle.rdbuf()->pubsetbuf(nullptr, 0); // no bufferring in fstream
    }

4KB 缓冲区大小不会重现问题。

Running main() from gtest_main.cc
Note: Google Test filter = *file_conc*_*append*
[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from fstream
[ RUN      ] fstream.file_concurrent_appends
file6.log Data written : 1 GB, 1 threads , cache true , size 512 bytes Time taken: 38069289 micro-secs
d:\projs\logpoc\tests\test.cpp(279): error:       Expected: num_records_each_thread * num_threads * offset
      Which is: 1073741824
To be equal to: file.tellg()
      Which is: 1073737728
d:\projs\logpoc\tests\test.cpp(301): error:       Expected: num_records_each_thread * num_threads
      Which is: 2097152
To be equal to: i + 1
      Which is: 2097145

编辑 2： 在加入所有线程以从内部缓冲区刷新数据后关闭file_handle。这解决了上述问题。

Answer 1

根据§29.4.2 ¶7 of the official ISO C++20 standard，std::fstream提供的函数一般都是线程安全的。

但是，如果每个线程都有自己的std::fstream 对象，那么就 C++ 标准库而言，这些是不同的流，不会发生同步。只有操作系统的kernel 会知道所有文件句柄都指向同一个文件。因此，任何同步都必须由内核完成。但是内核可能甚至不知道写入应该到达文件的末尾。根据您的平台，内核可能只接收特定文件位置的写入请求。如果文件的末尾同时被另一个线程的追加移动了，那么线程之前的写请求的位置可能不再是文件的末尾。

根据std::fstream::open 上的文档，以附加模式打开文件将导致流在每次写入之前查找文件末尾。这种行为似乎正是您想要的。但是，由于上述原因，这可能仅在所有线程共享相同的std::fstream 对象时才有效。在这种情况下，std::fstream 对象应该能够同步所有写入。特别是，它应该能够以原子方式执行对文件末尾的查找和后续写入。