· ☕ 4 分钟

https://tenzir.com/blog/production-debugging-bpftrace-uprobes/
https://shaharmike.com/cpp/vtable-part1/

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
#include <iostream>

class Parent {
 public:
  virtual void Foo() {}
  virtual void FooNotOverridden() {}
};

class Derived : public Parent {
 public:
  void Foo() override {}
};

int main() {
  Parent p1, p2;
  Derived d1, d2;

  std::cout << "done" << std::endl;
}
$ # compile our code with debug symbols and start debugging using gdb
$ clang++ -std=c++14 -stdlib=libc++ -g main.cpp && gdb ./a.out
...
(gdb) # ask gdb to automatically demangle C++ symbols
(gdb) set print asm-demangle on
(gdb) set print demangle on
(gdb) # set breakpoint at main
(gdb) b main
Breakpoint 1 at 0x4009ac: file main.cpp, line 15.
(gdb) run
Starting program: /home/shmike/cpp/a.out

Breakpoint 1, main () at main.cpp:15
15	  Parent p1, p2;
(gdb) # skip to next line
(gdb) n
16	  Derived d1, d2;
(gdb) # skip to next line
(gdb) n
18	  std::cout << "done" << std::endl;
(gdb) # print p1, p2, d1, d2 - we'll talk about what the output means soon
(gdb) p p1
$1 = {_vptr$Parent = 0x400bb8 <vtable for Parent+16>}
(gdb) p p2
$2 = {_vptr$Parent = 0x400bb8 <vtable for Parent+16>}
(gdb) p d1
$3 = {<Parent> = {_vptr$Parent = 0x400b50 <vtable for Derived+16>}, <No data fields>}
(gdb) p d2
$4 = {<Parent> = {_vptr$Parent = 0x400b50 <vtable for Derived+16>}, <No data fields>}

Here’s what we learned from the above:


· ☕ 3 分钟

Terminology

  • Cluster: a logical service with a set of endpoints that Envoy forwards requests to.

  • Downstream: an entity connecting to Envoy. This may be a local application (in a sidecar model) or a network node. In non-sidecar models, this is a remote client.

  • Endpoints: network nodes that implement a logical service. They are grouped into clusters. Endpoints in a cluster are upstream of an Envoy proxy.

  • Filter: a module in the connection or request processing pipeline providing some aspect of request handling. An analogy from Unix is the composition of small utilities (filters) with Unix pipes (filter chains).


· ☕ 0 分钟

· ☕ 1 分钟

https://istio.cn/t/topic/299

Envoy源码分析之Dispatcher:https://developer.aliyun.com/article/659277

线程相关 Misc:

  • Envoy进程由一个Main Thread和多个Worker Thread 组成
  • 每个Main和Worker包含一个eventloop,所有的处理都是由eventloop触发开始
  • Main负责xDS等功能,Worker负责处理连接和请求
  • 当一个client向Envoy建立连接的时候,因为所有Worker的EventLoop都注册了listening fd(启用SO_PORTREUSE除外),会由内核决定分配给哪个Worker
  • 当一个下游client连接到了Envoy,在保持连接不断的情况下,会和同一个Worker进行通讯

· ☕ 2 分钟

HTTP/1.1 Header Casing

https://www.envoyproxy.io/docs/envoy/latest/configuration/http/http_conn_man/header_casing#config-http-conn-man-header-casing

When handling HTTP/1.1, Envoy will normalize the header keys to be all lowercase. While this is compliant with the HTTP/1.1 spec, in practice this can result in issues when migrating existing systems that might rely on specific header casing.

To support these use cases, Envoy allows configuring a formatting scheme for the headers, which will have Envoy transform the header keys during serialization.

  • To configure this formatting on response headers, specify the format in the http_protocol_options.
  • To configure this for upstream request headers, specify the formatting in http_protocol_options in the cluster’s extension_protocol_options.

Currently Envoy supports two mutually exclusive types of header key formatters:


· ☕ 2 分钟

config.core.v3.Http1ProtocolOptions

https://www.envoyproxy.io/docs/envoy/latest/api-v3/config/core/v3/protocol.proto#config-core-v3-http1protocoloptions

[config.core.v3.Http1ProtocolOptions proto]

1
2
3
4
5
6
7
8
9
{
  "allow_absolute_url": "{...}",
  "accept_http_10": "...",
  "default_host_for_http_10": "...",
  "header_key_format": "{...}",
  "enable_trailers": "...",
  "allow_chunked_length": "...",
  "override_stream_error_on_invalid_http_message": "{...}"
}
  • allow_absolute_url

    (BoolValue) Handle HTTP requests with absolute URLs in the requests. These requests are generally sent by clients to forward/explicit proxies. This allows clients to configure envoy as their HTTP proxy. In Unix, for example, this is typically done by setting the http_proxy environment variable.


· ☕ 1 分钟

Enable debug log by command line

https://projectcontour.io/docs/v1.10.0/troubleshooting/envoy-debug-log/

The envoy command has a --log-level flag that can be useful for debugging. By default, it’s set to info. To change it to debug, edit the envoy DaemonSet in the projectcontour namespace and replace the --log-level info flag with --log-level debug. Setting the Envoy log level to debug can be particilarly useful for debugging TLS connection failures.

Enable debug log by API

列出 logger 名字:


· ☕ 2 分钟

https://www.envoyproxy.io/docs/envoy/latest/operations/performance

Performance

Envoy is architected to optimize scalability and resource utilization by running an event loop on a small number of threads. The “main” thread is responsible for control plane processing, and each “worker” thread handles a portion of the data plane processing. Envoy exposes two statistics to monitor performance of the event loops on all these threads.

  • Loop duration: Some amount of processing is done on each iteration of the event loop. This amount will naturally vary with changes in load. However, if one or more threads have an unusually long-tailed loop duration, it may indicate a performance issue. For example, work might not be distributed fairly across the worker threads, or there may be a long blocking operation in an extension that’s impeding progress.


· ☕ 3 分钟

Kernel

https://lwn.net/Articles/542629/
One of the features merged in the 3.9 development cycle was TCP and UDP support for the SO_REUSEPORT socket option; that support was implemented in a series of patches by Tom Herbert. The new socket option allows multiple sockets on the same host to bind to the same port, and is intended to improve the performance of multithreaded network server applications running on top of multicore systems.
The basic concept of SO_REUSEPORT is simple enough. Multiple servers (processes or threads) can bind to the same port if they each set the option as follows:


· ☕ 2 分钟

https://www.envoyproxy.io/docs/envoy/latest/intro/arch_overview/upstream/connection_pooling#automatic-protocol-selection

Connection pooling

For HTTP traffic, Envoy supports abstract connection pools that are layered on top of the underlying wire protocol (HTTP/1.1, HTTP/2, HTTP/3). The utilizing filter code does not need to be aware of whether the underlying protocol supports true multiplexing or not. In practice the underlying implementations have the following high level properties:

HTTP/1.1

The HTTP/1.1 connection pool acquires connections as needed to an upstream host (up to the circuit breaking limit). Requests are bound to connections as they become available, either because a connection is done processing a previous request or because a new connection is ready to receive its first request. The HTTP/1.1 connection pool does not make use of pipelining so that only a single downstream request must be reset if the upstream connection is severed.