added keepalive for dynomite listen connection

[TysonHeart]

Issue #691 Enabled keepalive for listen connection of dynomite.

[balajivenki]

@shailesh33 and @ipapapa Can you please review this PR for the ticket #691

[smukil]

@balajivenki Thanks for the contribution. I will have a look at it shortly.

[balajivenki]

@shailesh33 and @ipapapa Can you please review the changes? It will help us if it gets merged in the future releases.

[smukil]

@balajivenki @TysonHeart The code looks good to me and is doing the right thing. One additional thing I would note is whether you want to add an option for TCP_KEEPINTVL. The default on most linux kernels for the interval between 2 keep alive probes is 7200 seconds (2 hours), which I think is a fair interval for the purpose of recycling connections for our use cases. Setting the interval too less can add a lot of load to the server, especially when you're dealing with 50k+ connections, and leaving it too large can leave a lot of un-recycled connections.

If you'd like to add a configurable option for TCP_KEEPINTVL for your use cases, feel free to submit another patch. But this can go in as is for now. Thanks for the contribution again.

https://man7.org/linux/man-pages/man7/tcp.7.html
tcp_keepalive_time` (integer; default: 7200; since Linux 2.2)
The number of seconds a connection needs to be idle before TCP begins sending out keep-alive probes. Keep-alives are sent only when the SO_KEEPALIVE socket option is enabled. The default value is 7200 seconds (2 hours). An idle connection is terminated after approximately an additional 11 minutes (9 probes an interval of 75 seconds apart) when keep-alive is enabled.

[balajivenki]

@smukil Thanks for merging this. We thought of making it as configurable and decided to honor the values in sysctl.conf

net.ipv4.tcp_keepalive_time
net.ipv4.tcp_keepalive_intvl
net.ipv4.tcp_keepalive_probes

And we have changed our sysctl.conf to our ideal values. That way there is no confusion and all conf's related to tcp keepalive are at the same place.

[murthy-chelankuri]

Can we merge this change to our latest stable version as well v0.7.0 may be as v0.7.1? We are also facing a similar issue in our prod environment