asked on

Global Catalog servers and Exchange2000 server

I have 2 global catalog servers running on the same domain. 2 GC servers are set up for the purpose of fault tolerance. 1 Global catalog server is on the same segment and the same side of the router as the Exchange 2000 server. The other GC server is on a different segment and different side of the router.

The exchange 2000 server is set to use both GC servers. Whichever GC server was bootup 1st on the network, that server would be the primary GC server... Can you varify if the statement is correct.

Whenever 1 global catalog server goes down (for whatever reason), it affects Exchange (users can't get their email). Then we have to reboot the 2 Global Catalog servers and the Exchange 2000 server. I can resolve both GC's. Can be ping, name resolution and DNS are all functioning.

*****Questions*****
1. How come when 1 global catalog server goes offline, Exchange server 2000 won't recognize that there is still another global catalog server on line? Shouldn't Exchange server pick up the other GC seamlessly? What should I do to correct this?

2. Does it matter if is not on the same segment or side of router? If it does, why?

3.Should a GC server only be a domain controller and Global catalog server? Or could it have other functions?

Thanks for your help!

ASKER CERTIFIED SOLUTION

varoujB

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

David Wilhoit

Are you on SP2 or SP3 of Exchange>?

SOLUTION

metraton

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

metraton

I agree in part with varoujB, but i think that things run different in exchange, and also vary if you have one version or other of outlook, the first thing is that (according to MS) when a outlook client starts it contacts the EXCHANGE SERVER about what gc use to log, if the exchange isn't a gc it send a referral to the computer with the gc fqdn , and (in ol2000 till sr1) a persistent registry key points to the gc every time the client starts. If you take down that gc you must restart the client, then the client contacts exchange again and exchange gives another referral to another gc, or if he don't know about other then he fails. Yo must have a usable dns in the network, of coz, if you shutdown the dns in the network you can't logon correctly, only with cached creds. Another important point is that if exchange is also a gc, it dont use nspi or referrals and points himself as a gc. You can load balance gc's but with a key in the registry. If you want to know what you ol2000 and ol2000sr1 clients are using look at:

HKEY_CURRENT_USER\Software\Microsoft\Windows NT\CurrentVersion\Windows Messaging Subsystem\Profiles\Profile Name\dca740c8c042101ab4b908002b2fe182

in ol2000sr2 and forward all clients no hardcode the gc in the registry and asks for a gc every time they log. The only cache exchange have about gc's is in the dsaccess component.

Remember at least one dns avaliable with the correct registers in your network, one global catalog active and registered , and restart the clients (sometimes twice), they must be redirected by the exchange server to the alive gc.

Hope this helps

What90

No comment has been added to this question in more than 21 days, so it is now classified as abandoned.

I will leave the following recommendation for this question in the Cleanup topic area:
Split between metraton and varoujB

Any objections should be posted here in the next 4 days. After that time, the question will be closed.

What90
EE Cleanup Volunteer