I just sent the following e-mail to two mailing lists of former employees of a company I once worked at doing support, many of whom were supporting Exchange or NT. Figured it couldn’t hurt to cross-post it here as well.
Once upon a time, I couldn’t install NT4 on a new mail server, so I used Windows 2000 Server, setting up Exchange 5.5 on that.
I then started to follow directions for moving mailboxes, using test cases (a couple former employees and test accounts), which would ultimately have led to everything being moved and then changing which was the primary server for the site per available directions.
Trouble was, when I tested from Outlook 98 on an NT box, I got inability to resolve names and was at a loss. But duh, I never tried it from another machine, and when I got back to it, I found it was a problem with the NT box specifically. Weird.
So I did the ultimate test. I moved my mailbox, which was the largest (turns out it’s the second largest now).
Everything worked fine. I could connect. A copy of each e-mail still copied to this mail account.
Well, mostly. I started going long periods wondering what was up, sending myself tests, then getting them in bunches later. It looks like that is a result of RPC binding errors, don’t have the exact text from the event log on the proxy server/gateway with me, when the proxy server tries to talk to the new mail server.
Thoughts? (The next problem I mention may be related so hang on…)
What I gather based on an initial search is that it could be on account of network traffic, on account of past damage from a virus/worm (the proxy was never the same in subtle ways after it got Nimda way back when, and I have yet to reinstall the whole thing), or because it would be useful to go into raw mode – something I have never done or contemplated – and change a setting. None of what I saw offhand addressed the exact situation: Exchange 5.5 to Exchange 5.5, NT4 to Windows 2000. It seemed to be all about mixes of Exchange 5.5 and Exchange 2000.
Bear in mind that I am no Exchange admin. I simply do that as part of doing everything and somehow muddle through. The answer could be something simple.
The other thing, which I was in no position to discover until I moved my account, is that OWA is broken now.
When trying to use OWA for an account homed on the Windows 2000 server, it times out with:
error ‘ASP 0113′
Script timed out
I believe that is tied to at least some of the RPC errors I am seeing in the NT4 proxy’s event log, though perhaps those are actually all being generated when transferring mail from the NT4 to 2000 machine. Heck, for all I know all I have to do on 2000 is something that makes IMC talk to it properly.
As for OWA, any thoughts on the cause, and might the sometime mail delays relate to the OWA timeouts?
One more data point… There have been more general problems with the proxy lately. It’s been slow, more than I think any increase in traffic through it should justify. It’s needed more regular rebooting to get internet mail flowing again. It has Sybari Antigen and Spam Manager (well, they’re homed on the primary, to-be-replaced mail server, but Spam Manager does its work on the proxy) on it, and Spam Manager sometimes seems to strangle things, or simply stops working (which I think I have resolved by not letting its logs get too full). The machine has complained it was running out of space on the D drive with as much as 1.8 GB of free space. It was also hosting 9 GB of documents, but the speed and reliability problem remained after I moved the docs elsewhere. Figured I’d mention this in case the problem owes more to whatever ails that particular server than to any communication issue or setting between or on the Exchange 5.5/NT4 gateway and Exchange 5.5/Windows 2000 servers.
Any thoughts? I mean, besides getting the client to upgrade everything.