如何在Linux上验证TCP保持活动状态 [英] How to verify TCP keep alive on Linux

问题描述

我想在我的Linux机器上设置tcp keep live.所以我正在做的是运行脚本

I want to set tcp keep alive on my linux machine. So what I am doing is running a script

if [ `/sbin/sysctl -n net.ipv4.tcp_keepalive_time` != 200 ] ; then
   /sbin/sysctl -w net.ipv4.tcp_keepalive_time=200;

但是我仍然与亚马逊的红移有关.有人可以帮忙给我看看，我如何检查tcp keep live是否已设置?

But I still have issues with connections to amazon's redshift. Can someone please help and show me how I can check if tcp keep alive is actually set or not?

推荐答案

要检查保持活动状态是否为活动，请打开连接，请勿交换任何数据，并使用tcpdump或类似方法验证数据包是否在连接中定期交换，甚至如果没有实际数据交换.

To check if keep alive is active open a connection, don't exchange any data and verify with tcpdump or similar that packets gets regularly exchanged on the connection, even if no real data get exchanged.

但是也许您只是将keep-alive与超时混淆了:

But maybe you just confuse keep-alive with timeout:

• 保持活动状态:检测连接是否仍处于活动状态，例如必要时可以交换数据.这将检测到网络连接等的丢失.保持活动在TCP层上完成.
• 超时:一些不活动后，连接将关闭.这是在应用程序层完成的.

在两种情况下，除非尝试在套接字上进行读写，否则您都不会检测到问题.

In both cases you will not detect the problem until you try to read or write on the socket.

这篇关于如何在Linux上验证TCP保持活动状态的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！