Yea, when you find something out let me know. :(! I have exactly the
same problem, I can't get pkcipe or ciped-cb to do anything, or hint at
me what it is failing at. The documentation is long on theory and
configuration but there is not a "this is how to make it 'go'" anywhere.
OTOH, people are running it, so it must work.
I've been digging into the source, because I think optimally this type
of service should be of the client/server type where on one machine you
run a server that listens on a socket and can spawn as many interfaces
as neccessary for each connecting client.
This is the way I currently use 'ppptcp', but after utter failure to
grok the "firing-up" of cipe, I'm thinking of another re-write of ppptcp
stealing some of the ideas, maybe just running PPP on a UDP layer
connection (I can call it pppudp!). Frankly the encryption and keyring
is much simpler to implement with ppptcp.
Overall though I think the framework of CIPE is the optimal solution.
It should be able to obtain the efficiencies of NFS, by optimising out
all the inefficiencies of the TCP layer at the expense of writing more code.
Chris Wilson wrote:
> I'm trying to install cipe-1.5.2 on a linux 2.4.17 machine.
> I've compiled and installed the software... and created a config file:
> radar:/root # cat /etc/cipe/options
> ipaddr 18.104.22.168
> ptpaddr 22.214.171.124
> mask 255.255.255.255
> me 126.96.36.199:9999
> peer 188.8.131.52:9999
> key 1227d47e56766e989e43dfe4e6da1c56
> and with a similar mirrored config at the other end. I've loaded the
> kernel module - dmesg shows:
> cipcb: CIPE driver vers 1.5.2 (c) Olaf Titz 1996-2000, 100 channels
> cipcb: rtnl_lock() at ../cipe/device.c:625
> cipcb: rtnl_unlock() at ../cipe/device.c:627
> I've used the ip-up and ip-down scripts from the examples dir of the
> When I try to run ciped-cb either the program quits immediately or hangs
> with a child in the zombie state.
> No messages... no network traffic... no nothing....
> There doesn't appear to be any documention on the use of ciped (other than
> the -o option) but peeking at the source, nothing appears to be relevant.
> Any suggestions?
Keith Smith keith,AT,ksmith,DOT,com
655 W Fremont Dr
Tempe AZ 85282 it's hot