[OpenVPN home] [Date Prev] [Date Index] [Date Next]
[OpenVPN mailing lists] [Thread Prev] [Thread Index] [Thread Next]
Web openvpn.net

[Openvpn-users] 2.0b2: problematic `--ping' timer drift

  • Subject: [Openvpn-users] 2.0b2: problematic `--ping' timer drift
  • From: Joshua Judson Rosen <rozzin@xxxxxxxxxxxxx>
  • Date: Wed, 02 Jun 2004 15:08:17 -0400
  • Cancel-lock: sha1:RSI1nILnhxBNJppZ6JebUWaHQt8=

I've been experimenting with OpenVPN in scalable-udp-server mode using
two Debian GNU/Linux machines; after leaving the VPN running
overnight, I found the log showing *very* frequent ping-restarts: the
client was configured with "--ping-restart 60" and had got into a
state in which it was triggering restarts *every minute*, despite both
the server *and* the client being configured with `--ping 1'. The two
machines are the sole occupants of a new ethernet switch, so I figured
that the problem probably wasn't the network losing or abusing

After upping the verbosity on both the server and client, I can see
that both machines get into states in which they stop sending pings
for extended periods, and I can reproduce it--with `--verb 5', I can
conjure up a normal-looking `rwRWrRwWRWrw' output-string followed by a
long string of straight "W" or "R" characters--very disturbing. It
seems to correlate with the amount or rate (I'm not sure which) of
traffic flowing over the tunnel before stopping, e.g.: if I flood-ping
across the tunnel, the OpenVPN-pings go silent for a *long* time after
I've stopped my flood-ping.

I was hoping to be able to generate a patch for you, but I don't yet
understand the mechanics of the situation--any ideas what's going on
to make this happen? At which code should I be looking?

"It is easier to write an incorrect program
 than understand a correct one." --Alan Perlis

Openvpn-users mailing list