You might have noticed that I still had trouble believing that the packets are making it to the loghost, despite your 'netstat -a'. Just run 'netstat -t' on linux to see the TCP connections and see if you see the other box actually connected. Try the telnet test that I mentioned in the earlier post. Cover the lower layers of the stack before you concentrate on the app layer.
That was exactly my problem, there were no packages going anywhere, neither locally or remote. However if syslog-ng is runnin smoothly on the linux side of things, I assume tcp-wrappers has been installed and configured correctly. If the hosts.allow entry is correct then there's no reason it shouldnt accept the logs Anyway, I have now solved the problem. I tried installing 1.6.0rc1 (on solaris) however it wouldnt accept the fact that libol-0.3.9 really was that version. So I tried 1.5.26 and it runs smoothly, which 1.5.25 never did. next thing was on the linux side, I had to downgrade from a _very_ buggy 1.6.0_rc1-r1 syslog-ng gentoo release to (a still unstable) 1.5.26-r1. Now things seems to be working the way they should...logging locally and remotely seems to be working. Syslog-ng = a very touchy, senstive, but capable logger (once its runnin) cheers -AlanCB