Closing UTP server takes a long time

[CMCDragonkai]

After setting up a server and performing 1 successful connection with it, as soon as that connection is ended and destroyed, I try to close the UTP server, and it just hangs for about 20 to 25 seconds before it finishes.

I may have some UTP packets in the send buffer, but what are the expected amount of time taken for shutting down the UTP server since all connections have been destroyed.

[CMCDragonkai]

And this isn't really mentioned in the docs. But if I close the server while there are on-going connections, does the server automatically trigger the destruction of those connections? Or is just going to hang forever?

If so, I will need to go to each connection and destroy them right?

[mafintosh]

Yea you need to destroy each socket yourself similarly to a tcp server in node

[CMCDragonkai]

Every connection has been destroyed before I try to close the utp server. Is there some special trick to do so to hurry up the closing of the utp server?

…

On 13 May 2021 2:40:22 am AEST, Mathias Buus ***@***.***> wrote: Yea you need to destroy each socket yourself similarly to a tcp server in node -- You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: #40 (comment)

-- Sent from my Android device with K-9 Mail. Please excuse my brevity.

[mafintosh]

Can you share an example perhaps?

…

On 13 May 2021, at 02.25, Roger Qiu ***@***.***> wrote:  Every connection has been destroyed before I try to close the utp server. Is there some special trick to do so to hurry up the closing of the utp server? On 13 May 2021 2:40:22 am AEST, Mathias Buus ***@***.***> wrote: >Yea you need to destroy each socket yourself similarly to a tcp server >in node > >-- >You are receiving this because you authored the thread. >Reply to this email directly or view it on GitHub: >#40 (comment) -- Sent from my Android device with K-9 Mail. Please excuse my brevity. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[CMCDragonkai]

Well it's in a particular test case in our project js-polykey. Specifically this test case: https://github.com/MatrixAI/js-polykey/blob/nodes-refactoring/tests/network/ForwardProxy.test.ts#L138

[CMCDragonkai]

In my tests I'm creating 2 UTP sockets, one as the client socket and the other as the UTP server.

I've found that if I destroy all UTP connections, and then immediately close the client socket, this causes problems in closing the UTP server.

But if I wait 5 seconds before closing the client socket, the UTP server closes fine.

These tests are all running within 1 process.

Perhaps there's some issue with running multiple UTP sockets on the same process?

Or there is a problem if I were to close a UTP socket immediately after closing the the UTP connections?

[mafintosh]

PR a test case, and we'll take a look :)

[CMCDragonkai]

I've currently resolved this by doing:

utpSocket.close()
utpSocket.unref()

That way I'm not waiting for the socket to finish.

Also prior to this, I have to close all connections and also wait 1 second for utp to flush any buffers.