From: sam <sa...@hg...> - 2004-06-15 04:19:33
|
James Yonan wrote: >Adam Pavelec <apa...@be...> said: > > > >>On Monday, June 14, 2004 8:20 PM [GMT-5=EST], James Yonan <ji...@yo...> >>wrote: >> >> >> >>>>Which attributes of the certs need to be the same; and which >>>>attributes need to be different? >>>> >>>>I initially created certs for both the server and client with the >>>>exact same parameters and received an error while creating the >>>>client cert (TXT_DB error number 2), apparently because its >>>>attributes were identical to the server's cert. >>>> >>>>Then I created a server key with a commonName of something like >>>>'OpenVPN-Server' and then used 'OpenVPN-Client' for the client's >>>>cert. >>>> >>>>When I test the connection using the custom-tailored sample configs >>>>from the release notes, the TLS handshake fails, and the server >>>>spits out "Error: Windows resource limit WSA_MAXIMUM_WAIT_EVENTS >>>>(64) has been exceeded" errors while the client reports "TLS Error: >>>>Unroutable control packet received from ip.address.of.server:port >>>>(si=3 op=P_CONTROL_V1)". >>>> >>>> >>>The WSA_MAXIMUM_WAIT_EVENTS error is a real bug in beta4 and has been >>>fixed in beta5. >>> >>> >>Thanks for the info - I will move up to the current version. And please >>forgive my ignorance and inexperience, but am I creating my certs correctly? >>It looks as though my TLS handshake issue will still be problematic on >>beta5. >> >> > >The "TLS handshake failed" condition is very general and could simply mean >that there's no OpenVPN running on the other side of the connection, or >there's a link failure. > >If the sample certs provided work but your own certs don't, then I would >question whether or not your own certs were generated correctly. > >If the sample certs don't work, then it may just be a network connection issue. > >James > > > > I found a problem with this error was when client established connection with server, client only able to send initial requests to server, but never get packet back from the server. At client, I used tcpdump watch network traffic on tap device. I was never able to solve this problem, then I decided change the server platform to FreeBSD, because my orginal Redhat 9.1 server have vmware installed with bridge mode configured on other virtual network devices. I thought this cuased the routing problem with OpenVPN server. I don't meant OpenVPN does it incorrectly, but that may be base system caused incorrect routing. Sorry I don't have evident to prove what I m guessing here. Anyway, I have changed platform to FreeBSD, and now everything is working like a charm. sam |