I've done a small amount of testing with some of this a while back. I run a Samba server on each subnet and sync them through the tunnel(s), which has worked fine for WINS resolution for Win9x/ME machines. Win2K/XP is a completely different story however. What little I figured out with the browsing problems have been traced back to M$ breaking several RFC's with Name-resolution, so they could integrate WINS and DNS. This move breaks much of the compatibility with Samba and any other smb client/server OS other than Win2K/XP. I had some success by mapping drives (as others have noted), but I decided to drop Win2K/XP on the few boxes that where loaded with these OS's.
With M$ smb, you will need to get rid of any NetBUEI protocol instances since it is non-routable across subnets and use NetBIOS over TCP/IP.... the functionality is basically the same with cross- subnet access built in. Using hosts or lmhosts should be fine, but WINS is also acceptable... broadcast is a nightmare to use for name resolution. If using Samba (which I do suggest), make sure that the available subnets are allowed and the necessary server sync'ing is setup. All in all, mapping drives, as Charles' noted, is the best method performance wise in any case. Your luck may vary with any other method. I hope this helps!! On Friday 19 April 2002 17:08, Charles Steinkuehler wrote: > > I may have been one of those who replied on the FreeS/WAN list. > > Your posting has actually prompted me to revisit the whole issue. > > In brief, I think I said that the transfer speeds were fine so long > > as WINS and browsing was left out of the equation. At least that > > seems to be the case. However, as you know, this precludes using > > network neighbourhood. > > > > Do you need free run of network neighbourhood, or could you get by > > with several mapped drives? These could be done automagically with > > a logon script. > > I see *GIGANTIC* differences in access speed when browsing to network > resources vs using mapped drives. Mapping a drive makes the > performance *MUCH* better. I have yet to do any qualitative tests, > but "user experience" indicates a 10x to 100x type of improvement. > > This could easily be the result of a sub-optimal network > configuration on my part (I'm not a particularly good Windows admin, > and don't particularly want to be...I know just enough to keep my > system running). -- ~Lynn Avants aka Guitarlynn guitarlynn at users.sourceforge.net http://leaf.sourceforge.net If linux isn't the answer, you've probably got the wrong question! _______________________________________________ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
