Re: [Openvpn-devel] OpenVPN 2.0 Yo-Yo effect...

[James Yonan]

Denis Vlasenko  said:

> On Saturday 05 June 2004 21:46, Mike Auty wrote:
> > Thanks for the super fast reply,
> > Sadly I don't have the facilities to build a new windows version.
> > Would you mind exaplaining briefly what it does?  It only seems to
> > change the order of two checks, but I'm not sure what they do.  Thank
> > again, Mike  5:)
> 
> If there happens to be a packet to send when openvpn checks whether
> it needs to send a ping, openvps erroneously decide to set ping
> interval timeout to 365 days.

This is a known bug in 2.0-beta2 and will be fixed in beta3 which I expect to
release within a day or two.

James

Re: [Openvpn-devel] OpenVPN 2.0 Yo-Yo effect...

[Denis Vlasenko]

On Saturday 05 June 2004 15:36, Mike Auty wrote:
> Hi,
>   Since I've upgraded to OpenVPN 2, I've started seeing an interesting
> effect.  Originally (way back at _test23) it didn't happen very much,
> and seemingly a restart of either the server or client would fix it, but
> recently it's been happening a lot.
>   For historical reasons (I used to be on windows) I set my ping and
> ping-reset times to very low values (10 and 60) so that I'd see the
> adaptor go down within a minute of when my connection went down.  Under
> the old 1.5 vpns this was fine, but since the upgrade some of the
> windows installation's I've used have reported that the openvpn adaptor
> goes up and down every couple of minutes and in the recent _beta2 it
> seems to happen quite often (although it's still random in that
> sometimes it won't).

Try this patch
--
vda
--- ping-inline.h.orig	Wed Apr 21 08:03:22 2004
+++ ping-inline.h	Sun May 30 18:07:54 2004
@@ -50,9 +50,10 @@
 {
   void check_ping_send_dowork (struct context *c);
   if (c->options.ping_send_timeout
-  && !c->c2.to_link.len
   && event_timeout_trigger (>c2.ping_send_interval,
->c2.timeval))
+>c2.timeval)
+  && !c->c2.to_link.len
+)
 check_ping_send_dowork (c);
 }

[Openvpn-devel] OpenVPN 2.0 Yo-Yo effect...

[Mike Auty]

Hi,
	Since I've upgraded to OpenVPN 2, I've started seeing an interesting 
effect.  Originally (way back at _test23) it didn't happen very much, 
and seemingly a restart of either the server or client would fix it, but 
recently it's been happening a lot.
	For historical reasons (I used to be on windows) I set my ping and 
ping-reset times to very low values (10 and 60) so that I'd see the 
adaptor go down within a minute of when my connection went down.  Under 
the old 1.5 vpns this was fine, but since the upgrade some of the 
windows installation's I've used have reported that the openvpn adaptor 
goes up and down every couple of minutes and in the recent _beta2 it 
seems to happen quite often (although it's still random in that 
sometimes it won't).
	So far I've observed this between windows-windows and windows-linux 
machines.  I haven't seen it between linux and linux boxes, but that may 
be because it's harder to spot (since there isn't an adaptor up/down 
indicator by default).  Having checked some recent logs I haven't seen 
it happen much between linux-linux (just the hourly key-expiration).
	Unfortunately the closest example I have to hand doesn't have client 
ping-reset as well as server ping-reset.  The log for the client list 
says it's starting up ok, then one line saying replay window backtrack 
and then no other messages (for about half a day in fact), and the 
server reports inactivity timesout approximately every minute or so.  As 
far as I'm aware the internet connections are fine.
	So basically I was wondering if there was something in the ping code, 
or the inactivity checker that may have changed that someone knows about 
which could account for this behaviour?  I'll be happy to provide 
configs/logs if they'd be at all helpful...

Thanks,
Mike  5:)