Periodically drops Tx frames and requires disconnect/reconnect

[Aaron]

I use openconnect / network-manager-openconnect to connect the VPN for work. In Ubuntu 17.04, this worked perfectly fine and I would remain connected for hours on end, without fail.

Since upgrading to 17.10, however, at least once an hour all traffic suddenly halts and the network becomes unreachable. Running "ifconfig" from the terminal shows something like this:

vpn0: flags=4305<UP,POINTOPOINT,RUNNING,NOARP,MULTICAST> mtu 1406
        inet XXX.XXX.XXX.XXX netmask 255.255.0.0 destination XXX.XXX.XXX.XXX
        inet6 XXXX::XXXX:XXXX:XXXX:XXXX prefixlen 64 scopeid 0x20<link>
        unspec 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00 txqueuelen 500 (UNSPEC)
        RX packets 537872 bytes 385640344 (385.6 MB)
        RX errors 0 dropped 0 overruns 0 frame 0
        TX packets 426702 bytes 82214569 (82.2 MB)
        TX errors 0 dropped 7071 overruns 0 carrier 0 collisions 0

the "dropped" column of TX errors continuously increases when I try to make any outgoing requests. No incoming data comes in, at all.

I have not been able to isolate specifically what's causing this and I've not been able to reproduce it intentionally beyond just "connect to the VPN and use it until it breaks". Non-VPN internet works fine.

Connection is over a hardwired ethernet connection, onboard NIC. (Dell)

ProblemType: Bug
DistroRelease: Ubuntu 17.10
Package: openconnect 7.08-1
ProcVersionSignature: Ubuntu 4.13.0-16.19-generic 4.13.4
Uname: Linux 4.13.0-16-generic x86_64
ApportVersion: 2.20.7-0ubuntu3.1
Architecture: amd64
CurrentDesktop: ubuntu:GNOME
Date: Thu Nov 9 14:59:01 2017
InstallationDate: Installed on 2015-12-21 (689 days ago)
InstallationMedia: Ubuntu 15.10 "Wily Werewolf" - Release amd64 (20151021)
ProcEnviron:
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: openconnect
UpgradeStatus: Upgraded to artful on 2017-10-24 (16 days ago)

[Marc Dietrich (marvin24)]

I'm also affected. Looking at the process list (ps ax) I found:

23386 pts/4 S+ 0:00 openconnect -vvvvv -s /usr/share/vpnc-scripts/vpnc-script vpn.example.com
23400 pts/4 S+ 0:00 /bin/sh -c /usr/share/vpnc-scripts/vpnc-script
23405 pts/4 S+ 0:00 /bin/sh /usr/share/vpnc-scripts/vpnc-script
23440 ? Ssl 0:00 /usr/lib/NetworkManager/nm-dispatcher
23443 ? S 0:00 /bin/sh -e /etc/NetworkManager/dispatcher.d/01-ifupdown tun0 up
23468 pts/4 S+ 0:00 run-parts --arg=-a --arg=tun0 /etc/resolvconf/update.d
23479 pts/4 S+ 0:00 run-parts /etc/resolvconf/update-libc.d
23500 ? S 0:00 /bin/sh /etc/network/if-up.d/ntpdate
23502 ? S 0:00 flock -n /run/lock/ntpdate /usr/sbin/ntpdate-debian -s
23504 ? S< 0:00 /usr/sbin/ntpdate -s de.pool.ntp.org
23515 pts/4 S+ 0:00 /bin/sh /usr/lib/avahi/avahi-daemon-check-dns.sh
23534 pts/4 Sl+ 0:00 host -t soa local.
23539 ? S 0:00 run-parts /etc/network/if-up.d
23550 ? S 0:00 /bin/sh /usr/lib/avahi/avahi-daemon-check-dns.sh
23567 ? Sl 0:00 host -t soa local.
23574 pts/3 R+ 0:00 ps ax

the "avahi-daemon-check-dns.sh" process hangs, maybe because the route to dns isn't setup yet (vpnc-script still running). If I kill this process (pid 23550) the script continues to run and the connect is alive and stable.

[Marc Dietrich (marvin24)]

sorry, above comment was meant to be added to bug 1752411.