Client connection is lost with no obvious reason

[zllvm]

Client connection is lost with no obvious reason

[zllvm]

I've done the following change in PollSocket method:

catch (SocketException ex)
{
    Logger?.Invoke($"{Header}poll socket from {ServerIpPort} failed with ex = {ex}");
    return ex.SocketErrorCode == SocketError.TimedOut;
}

After that, I deployed it to the dev server and started a TCP connection that has been up and running for two days with no issues. I could test it by running various commands on the open connection, which worked perfectly. The log entry that I've added $"{Header}poll socket from {ServerIpPort} failed with ex = {ex}" appeared several time times (~20) in the monitoring system with Operation timed out exception.

Evidently, PollSocket introduces a harmful behavior by disconnecting an existing connection. I don't know if it is possible to make it fully reliable. If not, then it should be allowed to override or disable it.

What is your opinion on this, @jchristn? Could we re-open this issue?

[jchristn]

Do you have TCP keepalives enabled on either client or server? I don't suspect there is an issue in the implementation, but if you can point to one, I'd love a PR with a fix!

[zllvm]

I have keepalive enabled on the client but not on the test server. I've found another issue where you explain that keepalives can bring odd behavior. I can disable them and see if it helps.

It is worth mentioning that I cannot reproduce this issue on my windows pc, and it fails in the ci pipe that executes in a Linux environment (mcr.microsoft.com/dotnet/sdk:6.0-alpine). Therefore, it seems more likely to be something related to keepalives.

This issue can be closed, I suppose. And if I find a problem in the library code that I can fix, I'll do a PR as you suggest. Thanks!

[jchristn]

Hi @zllvm TCP keepalives are the bane of my existence (at least with these libraries...). I have unfortunately zero control over them, and they only exist to solve a very niche problem (cable removed from switch or router) and it often doesn't even do a good job of solving that, not to mention they seemingly introduce random problems everywhere. Please let me know if disabling them helps. Please re-open if the problem continues with keepalives disabled! Cheers, Joel

[zllvm]

Hi @jchristn! Aw, I understand your pain!

It works much better with keepalives disabled, though it still disconnects occasionally. I started to debug it by increasing logging in your library. It turns out that it fails on:

var clientSentData = _client.Client.Receive(buff, SocketFlags.Peek) != 0;

with System.Net.Sockets.SocketException (110): Operation timed out

After that, the connection is forcefully closed by ConnectedMonitor. Nevertheless, I'm pretty sure that the connection is still alive.

If I understand it correctly, Receive is called in a non-blocking mode and therefore throws SocketException when there's no data according to the microsoft doc. In this document, it's suggested to check the Available property to know if there's data.

Thereby, I realized that I don't fully comprehend the logic in PollSocket:

• If no data is available at a specific moment, how could Receive indicate a connection problem?
• Can we be sure that SocketException on Receive always implies a connection loss?

Likewise, what do you think about configuring the ConnectedMonitor task so it can be optionally disabled?

[Energiz0r]

FYI: We are having the exact same issue in our environment here.
Works stable when testing this on windows, but as soon as we deploy this to our Ubuntu server, disconnections start with Timeout 110 error.

[Espen-Kalhagen-Element-Logic]

FYI: We are having the exact same issue in our environment here. Works stable when testing this on windows, but as soon as we deploy this to our Ubuntu server, disconnections start with Timeout 110 error.

I work with Energiz0r.
We need to use SuperSimpleTcp with the changes suggested by @zllvm, returning true if the exception is a timeout exception, to achieve stable tcp connections.

[jchristn]

Hi, I will take a look today. Cheers

[jchristn]

Hi, could you please try v3.0.7?

NuGet: https://www.nuget.org/packages/SuperSimpleTcp/3.0.7
Commit: 825c017

[jchristn]

Please re-open if this doesn't address the issue. Cheers!