weird vpn routing problem

Hi all,

I am desperate for a little help. I have a server (RH 7.1,
2.4.2-2)with a 'receive-only' sat-link (using vpn, pptp) and a dsl
line (for sending) that is connected thru a cisco 1601 router.
I should mention that everything was working nicely and then something
happened.! and I don't know what.

When the sat-link is up everything is working except I can not telnet
to the server from the 'outside'. This also means I can not receive
mail. When the sat-link is down it reverts to using the dsl line for
sending and receiving, and when this happens everything is working
incl. receiving mail and telnetting into the server from the outside.

Right so far so good,

Here comes the weird part: when the sat-link is up (and thus mail not
working etc.) I can receive mail and telnet from (only) the isp that
is selling me the vpn-sat service. If I send myself a mail from yahoo
it never arrives (only when the satlink is down), and if I traceroute
from outside it can not reach the server (again only when satlink is
down).

However (when the satlink is up) I *can* telnet from outside to the
cisco router. And if I add a host route that points to the ip of the
'outside' computer then voilla... telnet works.

My routing table looks like this (slightly sanitized):

Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref
Use Iface
196.38.110.84 * 255.255.255.255 UH 0 0
0 eth0
196.38.110.84 * 255.255.255.255 UH 0 0
0 ppp0
196.28.aaa.248 196.28.238.249 255.255.255.252 UG 0 0
0 eth0
196.28.aaa.248 * 255.255.255.252 U 0 0
0 eth0
10.1.1.0 * 255.255.255.0 U 0 0
0 pentaval0
192.168.10.0 * 255.255.255.0 U 0 0
0 eth1
127.0.0.0 * 255.0.0.0 U 0 0
0 lo
default 196.38.110.84 0.0.0.0 UG 0 0
0 ppp0

My router has ip: 196.28.aaa.249
Server has: 196.28.aaa.250
netmask is: 255.255.255.252 as you can see.
196.38.110.84 is the tunnel peer

or,
#ip route list
196.38.110.84 dev eth0 scope link
196.38.110.84 dev ppp0 proto kernel scope link src 196.36.99.207
196.28.aaa.248/30 via 196.28.aaa.249 dev eth0
196.28.aaa.248/30 dev eth0 proto kernel scope link src
196.28.aaa.250
10.1.1.0/24 dev pentaval0 proto kernel scope link src 10.1.1.2
192.168.10.0/24 dev eth1 proto kernel scope link src 192.168.10.1
127.0.0.0/8 dev lo scope link
default via 196.38.110.84 dev ppp0

It is probably important to mention that the guy from the isp that
*can* send me mail and telnet is also on the 196.0.0.0 network.

As I mentioned above, everthing *was* working and then something
happened.
The server had a crash (power failure) around the time when this
problem happened, and I am wondering if that could have an effect? I
am using ext2 and It took a long time for the server to fix itself.
So any suggestions as to what I can check? what should I try? I have
stared at this problem for a week and have probably gone blind :-)

Regards,
Tobias Skytte