Show HN: Herd – 一个 Go 侧车，用于阻止有状态进程（Puppeteer/LLM）因 OOM 崩溃
2 分•作者: sankalpnarula•5 个月前
大家好，HN。<p>我是滑铁卢大学的一名工程系学生，正在构建有状态的 AI 代理，我一直遇到同样的问题：每当我的 Python 脚本崩溃或断开连接时，底层的 Puppeteer 或 Ollama 进程就会一直孤零零地运行，占用 RAM，直到节点因内存溢出而自我终止。标准的负载均衡器会破坏粘性会话，而被动的 HTTP 超时对于清理来说又太慢了。<p>我找不到一个好的本地进程池，能够真正可靠地清理已死的有状态会话，所以我用 Go 语言构建了 Herd。<p>它使用一个持久流（gRPC/Unix 套接字）严格地作为“死亡男人开关”。如果你的客户端脚本死亡，流就会中断。Herd 会注册 EOF 并立即向工作进程发送 SIGKILL 信号（依赖于 Linux 上的 Pdeathsig）。对于实际的繁重数据，你只需通过 Herd 的内部代理发送 HTTP 流量，它会将流量直接路由到活动进程端口。<p>我的实际目标是将其变成一个多节点分布式网格，带有 Redis 注册中心，客户端可以退出，边缘网关会将它们路由回持有其有状态内存的确切 pod。<p>但我知道在有漏洞的本地引擎之上构建分布式网格是死路一条。单节点清理必须首先做到完美。<p>我很乐意让你们来“吐槽”这个架构。具体来说：在生产环境中，依赖 Pdeathsig 作为本地“死亡男人开关”是否足够健壮，还是我太天真了，现在就得咬紧牙关，把所有东西都封装在 cgroups 和微型虚拟机中？<p>代码库链接：<a href="https://github.com/herd-core/herd" rel="nofollow">https://github.com/herd-core/herd</a>
查看原文
Hey HN.<p>I'm an engineering student at Waterloo building stateful AI agents, and I kept hitting the same wall: whenever my Python scripts crashed or dropped a connection, the underlying Puppeteer or Ollama processes would just sit there orphaned, eating RAM until the node OOM-killed itself. Standard load balancers break sticky sessions, and passive HTTP timeouts are too slow for cleanup.<p>I couldn't find a good local process pool that actually cleaned up dead stateful sessions reliably, so I built Herd in Go.<p>It uses a persistent stream (gRPC/Unix sockets) strictly as a dead-man's switch. If your client script dies, the stream breaks. Herd registers the EOF and instantly fires a SIGKILL to the worker process (relying on Pdeathsig on Linux). For the actual heavy data, you just blast HTTP traffic through Herd's internal proxy, which routes it directly to the active process port.<p>My actual goal is to turn this into a multi-node distributed mesh with a Redis registry, where a client can drop off and an edge gateway routes them back to the exact pod holding their stateful memory.<p>But I know building a distributed mesh on top of a leaky local engine is a death sentence. The single-node cleanup has to be flawless first.<p>I'd love for you guys to roast the architecture. Specifically: is relying on Pdeathsig actually robust enough for a local dead-man's switch in production, or am I being naive and need to just bite the bullet and wrap everything in cgroups & microvms right now?<p>Repo link: <a href="https://github.com/herd-core/herd" rel="nofollow">https://github.com/herd-core/herd</a>