diff --git a/i18n/en-us/docusaurus-plugin-content-docs/current/doc/performance.md b/i18n/en-us/docusaurus-plugin-content-docs/current/doc/performance.md index a7b8c602..41e89eba 100644 --- a/i18n/en-us/docusaurus-plugin-content-docs/current/doc/performance.md +++ b/i18n/en-us/docusaurus-plugin-content-docs/current/doc/performance.md @@ -284,6 +284,33 @@ query check=added > Note: To avoid interference from the HTTP request itself on Valgrind, SRS uses a separate coroutine to perform periodic checks. Therefore, after accessing the API, you may need to wait a few seconds for the detection to be triggered. +Sometimes, you will receive the `still reachable` report for static or global variables, like this example: + +```text +==3430715== 1,040 (+1,040) bytes in 1 (+1) blocks are still reachable in new loss record 797 of 836 +==3430715== at 0x4C3F963: calloc (vg_replace_malloc.c:1595) +==3430715== by 0x7D8DB0: SrsConfig::get_hls_vcodec(std::__cxx11::basic_string, std::allocator >) (srs_app_config.cpp:7156) +==3430715== by 0x781283: SrsHlsMuxer::segment_open() (srs_app_hls.cpp:418) + +# It's caused by static variable: +string SrsConfig::get_hls_vcodec(string vhost) { + SRS_STATIC string DEFAULT = "h264"; + SrsConfDirective* conf = get_hls(vhost); + if (!conf) { + return DEFAULT; +``` + +You can easily work around this by publishing the stream, stopping it, and then triggering the memory leak detection: + +1. Publish stream to initialize the static and global variables: `ffmpeg -re -i doc/source.flv -c copy -f flv rtmp://127.0.0.1/live/livestream` +1. Trigger memory detection by using curl to access the API and generate calibration data. There will still be many false positives, but these can be ignored: `curl http://127.0.0.1:1985/api/v1/valgrind?check=added` +1. Retry memory detection, util the valgrind leak summary is stable, no any new lost blocks. +1. Perform load testing or test the suspected leaking functionality, such as RTMP streaming: `ffmpeg -re -i doc/source.flv -c copy -f flv rtmp://127.0.0.1/live/livestream` +1. Stop streaming and wait for SRS to clean up the Source memory, approximately 30 seconds. +1. Perform incremental memory leak detection. The reported leaks will be very accurate at this point: `curl http://127.0.0.1:1985/api/v1/valgrind?check=added` + +With the variables initialized, the `still reachable` report will be gone. + ## Syscall Please use [strace -c -p PID](https://man7.org/linux/man-pages/man1/strace.1.html) for syscal performance issue. diff --git a/i18n/zh-cn/docusaurus-plugin-content-docs/current/doc/performance.md b/i18n/zh-cn/docusaurus-plugin-content-docs/current/doc/performance.md index 8a884999..3f97dffc 100644 --- a/i18n/zh-cn/docusaurus-plugin-content-docs/current/doc/performance.md +++ b/i18n/zh-cn/docusaurus-plugin-content-docs/current/doc/performance.md @@ -285,6 +285,33 @@ query check=added > Note: 为了避免HTTP请求本身对valgrind的干扰，SRS使用单独的coroutine定时检查，因此访问API后，可能需要等待几秒才会触发检测。 +有时候，你会收到很多`still reachable`的报告，这是因为静态或全局变量，比如这个例子： + +```text +==3430715== 1,040 (+1,040) bytes in 1 (+1) blocks are still reachable in new loss record 797 of 836 +==3430715== at 0x4C3F963: calloc (vg_replace_malloc.c:1595) +==3430715== by 0x7D8DB0: SrsConfig::get_hls_vcodec(std::__cxx11::basic_string, std::allocator >) (srs_app_config.cpp:7156) +==3430715== by 0x781283: SrsHlsMuxer::segment_open() (srs_app_hls.cpp:418) + +# It's caused by static variable: +string SrsConfig::get_hls_vcodec(string vhost) { + SRS_STATIC string DEFAULT = "h264"; + SrsConfDirective* conf = get_hls(vhost); + if (!conf) { + return DEFAULT; +``` + +你可以调整下测试步骤，在触发内存检测前，先推流来初始化这些静态和全局变量，然后再次推流和内存增量泄露的检测： + +1. 启动SRS后先推流，初始化静态和全局变量，然后停止推流: `ffmpeg -re -i doc/source.flv -c copy -f flv rtmp://127.0.0.1/live/livestream` +1. 触发内存检测，使用curl访问API，形成校准的数据，依然有大量误报，但可以忽略这些数据：`curl http://127.0.0.1:1985/api/v1/valgrind?check=added` +1. 多尝试几次内存检测，直到没有新增的泄露，包括possibly和reachable。 +1. 压测，或者针对怀疑泄露的功能测试，比如RTMP推流：`ffmpeg -re -i doc/source.flv -c copy -f flv rtmp://127.0.0.1/live/livestream` +1. 停止推流，等待SRS清理Source内存，大概等待30秒左右。 +1. 增量内存泄露检测，此时报告的泄露情况就非常准确了：`curl http://127.0.0.1:1985/api/v1/valgrind?check=added` + +由于变量已经初始化了，所以`still reachable`的干扰报告就不会有了。 + ## Syscall 系统调用的性能排查，参考[strace -c -p PID](https://man7.org/linux/man-pages/man1/strace.1.html)