<< | Thread Index | >> ]    [ << | Date Index | >> ]

Subject: Timeout problems
From: Al Smith <Al.Smith,AT,aeschi,DOT,ch,DOT,eu,DOT,org>
Date: Fri, 17 Aug 2001 10:07:10 +0200

Hi there,

I've been running CIPE 1.5.2 for a couple of months now, and am
encountering some intermittent problems.

Cipe is started using pkcipe; that works with no problems. The tunnel
comes up, ip-up adds the relevant routes on each end, and everything
works fine.

However when the remote end kills the daemon and then disconnects from the
internet, the server logs "peer closing connection". But occasionally the
ciped on the server doesn't actually die. Of course next time the client
starts up pkcipe then the server has then a problem with conflicting
routes and interfaces to the same remote address.

I should perhaps note that this is running on linux 2.2.19, and the same
failure mode is noted in both the cb and db versions of cipe.

The syslog notes "keepalive timeout", and after modification to display
the errcnt variable, it emerges that the errcnt is never successfully
decremented in between successive "keepalive timeout" log calls, and thus
never reaches zero.

Questions: should the ciped notice when the other side sends it "going
down" and then die automatically? And I also wonder why errcnt
consistently gets reset to maxerr during kxchg()?

In the config file for the connections, I have the following. Should this
be somehow changed?

cttl  64
maxerr 3
ping  20
toping 2

Many thanks,


<< | Thread Index | >> ]    [ << | Date Index | >> ]