Network recovery is not working well on a VPN and sleep setup

[spidey]

I use matterhorn on my work laptop. We have a VPN connection that is maintained automatically in the background. I also lock and sleep the laptop a lot just by closing the lid.

Running matterhorn 50200.17.0 (master@7ca52c1) with Mattermost API 50200.13.0 (develop@273adac) I'm usually not able to recover the connection and I'm relying more and more on quitting and reopening matterhorn. That is happening perhaps 10 times a day.

What can I do to diagnose the underlying networking issue to try and make this a feature request for more resilient networking code from matterhorn?

[jtdaugherty]

@spidey Thanks for opening this, and I'm sorry to hear that you are not having a good experience with Matterhorn's recovery behavior. I'm not exactly sure what to do here, because:

1. Matterhorn should automatically retry a connection if the current one is severed;
2. If Matterhorn experiences an unexpected network problem that it cannot handle, it should report that to you in the chat UI.

Presumably (2) is not happening because if it was, you'd get a message asking you to report it here. And clearly (1) is not quite working; my guess is that (1) is attempting to reconnect, but the network stack is so stuck that the attempted reconnect just hangs. You could enable logging in Matterhorn (via command-line flag or /log-start in the program) and send me a log for a timespan in which this issue occurs. I'm not sure it would show much, but it's a place to start. On the assumption that my hypothesis is correct, we could have Matterhorn also time out on its connection reattempt and try again, which could sidestep a stuck network stack (a common issue in VPN usage: routing tables and nameservers get changed and connection attempts made during that time can get into a stuck state).