MikroTik

Hello, I am using RB433AH to connect to PPTP running on FreeBSD. The server is well established and works with a variety of clients.

The RB initially had v3.25, but I moved to v3.30, and then v4.5. The problem across all versions is that the LCP negotiation does not seem to succeed. I have set the server to be as flexible as possible, but the RB seems to reject the asyncmap option. According to the ppp man page for the server, the asyncmap option is ALWAYS negotiated, with no option to disable it. You can try a different asyncmap value, but you can't stop the item from being negotiated. The feature is called accmap in the ppp man page.

Obviously, the connection attempt fails after a few LCP negotiation rounds.

Is there an asyncmap value that the RB wants, or is the RB just refusing asyncmap entirely, hence the never-connect situation?

I modified & compiled ppp to not send the accmap(asyncmap) item, and to just assume an asyncmap of 0 for both sides.

After turning off other items that Mikrotik was rejecting (afccomp, protocomp, vjcomp), and setting mppe 128 stateless, I am getting connections and passing traffic...for about six minutes.

After that, there seems to be some unknown glitch. After about 6 minutes, the traffic stops passing. PPTP echo and replies are still moving. After disable and enable of the PPTP interface, it still will not connect for MINUTES! The log indicates that the remote host is 'unreachable' as it repeatedly dials. I know this is not the case. Other hosts on the same ethernet can reach the same PPTP server with no problem.

SO, my conclusion is that there is some either a routing misconfigure on my part, or a bug on Mikrotik part. Any ideas?

Other random problems:

Mikrotik should properly negotiate the asyncmap by accepting the offer of 0. (And of course, FreeBSD ppp shouldn't insist on it either)

Sometimes a negotiation succeeds, other times it fails. I can only assume this is due to bugs in Mikrotik, because the server is long established and does not have this problems with a variety of clients.

Mikrotik logs the IPv6CP as an 0x8057 'unrecognized protocol' and doesn't respond (except with the unrecognized protocol message), slowing the negotiation process. Mikrotik should respond by rejecting IPv6CP if they don't support IPv6. I turned it off on the server side to avoid this.

After a PPTP session is up, it takes a minute or two before packets will flow through.

Mikrotik can't seem to handle LQR either. I turned that off on server side.

If the server doesn't insist on either stateless or stateful, the connection ends up in a combo of the two. Which doesn't seem to work. Not sure if this is a FreeBSD ppp bug, or Mikrotik.