[Pudding] Pudding https://forum.qubes-os.org/u/pudding
Hello, my setup is as follows:
Debian 10 AppVM (Chromium) - OpenVPN VMs (Mullvad or ProtonVPN, with
Qubes-VPN-support made by Chris Laprise) - sys-whonix - clearnet
Recently I found that whenever I use Chromium inside my AppVM, OpenVPN
always constantly reconnects. I looked up OpenVPN log and found these
messages appeared many times:
Connection reset, restarting 
TCP/UDP: Closing socket
SIGUSR1[soft,connection-reset] received, process restarting
Restart pause, 5 second(s)
Hi… That sounds really bizarre.
I just tried Chromium (Debian 10) over a Mullvad connection and it seems
to work fine. Openvpn is running from the recent update of
Qubes-vpn-support (1.4.4) in a Debian VM. The Qubes version is 4.1alpha
but I don’t think that matters here.
Have you tried Firefox as a test?
Both Mullvad and ProtonVPN VMs have this issue. I tried different
Chromium versions, namely official 83.0.4103.116 in Debian Stable repo,
ungoogled-chromium 84.0.4147.135-1.buster1 and
87.0.4280.67-1.unportable1, and Chromium 88.0.4287.0 AppImage. Only
official 83.0.4103.116 does not cause this issue. But I prefer
ungoogled-chromium for privacy reasons. Is there any way to fix this?
Thanks in advance.
I wish I had an idea off the top of my head; if I think of one I’ll post
One avenue to explore (though it seems a bit unlikely) is to find out if
the special-build Chromium is doing anything really unusual with DNS,
TLS or TCP/UDP protocols.
Another thing is you can increase the Openvpn log verbosity to something
like 5 to get more detail.
You can also try stopping the qubes-vpn-handler service and starting
openvpn manually with the config file in /rw/config/vpn. I suggest this
because the one thing that Qubes-vpn-support alters about the connection
parameters are the variables that control connection timeout, and