Exchange Server Unavailable

[MIScoord]

Having a really weird situation that I need some help on. We have a W2K server system, which is a DC that we're also running Exchange 2K on. Most of our Outlook clients are fine, but we have a few that are experiencing some strange difficulties. These clients can authenticate to the network without a problem, and they can resolve the Exchange Server name and mailbox name from within mail properties without a problem. However, when they start Outlook the receive a "Exchange Server unavailable. Do you want to work offline?" message. Now these clients are a mix of different machines -- some are older Pentium 200 machines, some are 400s, some with Outlook 98, some with Outlook 2000. However, the only common denominator is that they are all at satellite locations and not at the main site. In other words, their request has to travel from their machine to a router, through a frame relay to another router to the Exchange Server. This isn't a problem with authentication and name resolution, and it only happens with a few machines (the machine in one office will have problems and the one next door will work fine), so I'm relucatant to pronounce it a bandwith problem. Anyone have any ideas?

Me: We need a better backup system.
My boss's boss: Backup? We don't need no stinkin' backup!

 

[Zelandakh]

Do the satellite sites have DNS and WINS servers? If not, that is your problem. Create lmhosts and hosts files with correct entries, ip addresses, #domain tags and so on. Drop them out to the affected machines and it should solve the problem. Bandwidth to Exchange can be fairly minimal, finding the way there before it times out.

 

[MIScoord]

That's the kicker. The main location where this is happening DOES have a W2K domain contoller. It isn't where Exchange resides, but it should forward the request to the Exchange server without needing a host or lmhost file.

One thing I guess I should have mentioned in the previous entry: We did the upgrade on our system back in December, and when we were operating under NT 4.0, we ran DHCP on our routers and assigned IP addresses there. When we upgraded to W2K, we decided to move DHCP to our main server here, so we could better track stuff. However, the boot-p forwarding on our brand of router is a bear to figure out how to set up, and in order to work around the problem and get the servers up and running, we staticly assigned IP addresses to all the clients. About three weeks ago, we finally found the right documentation on the routers to set up boot-p, so we did it not only on the LAN router at the main location, but also on the routers at the satellites. However, we still aren't using DHCP for assigning IP addresses, because the routers don't seem to be forwarding the requests correctly at some of our locations. I'm beginning to wonder now if turning on boot-p on the routers may be causing the problems we're having with Outlook/Exchange. I don't see how, but at this point, I'm grasping at straws. Me: We need a better backup system.
My boss's boss: Backup? We don't need no stinkin' backup!