This post is more than 5 years old

9 Posts

593

June 22nd, 2007 08:00

Replistor 6.2 with Exchange 2003

The configuration is: One Exchange 2003 server at the primary site and we are setting up a new Exchange 2003 server at the DR (disaster recover) site.
We will be setting up Replistor to replicate data on the Exchange server between sites.

1. When you configure the Replistor alias for failover on an existing Exchange server how does that effect the current Exchange clients that are pointing to the real server name? Per the EMC Replistor Release Notes there should be no impact.

2. How will other Exchange dependent applications (such as Unity voicemail & Blackberry Enterprise server) work upon "failover" to the target Exchange (disaster recovery) server? The EMC Admin Guide document discusses Exchange failover but not other Exchange related services.

4 Posts

June 27th, 2007 16:00

For question 1) Client machines should be pointing at the Alias, not the real name of the production Exchange server, in turn the Alias points to the production which then gets the clients pointed at the real server. In a failover situation Replistor would modify the Alias in DNS to point to the DR Exchange server .. clients shouldn't notice.

For Question 2) The Blackberry and Unity servers should also be pointed at the Alias and things should then work the same way as the clients .. since thats all they really are, just a sort of "super client" :-) If you are running the Blackberry services on the same physical server as Exchange then you need to make a script that gets kicked off as part of the failover process which would then start up the Blackberry services on the DR server as well.

Basically if a client or server has to interact with the Exchnage server ... then ALWAYS have them referring to the Alias, not the "real" server name.

45 Posts

July 30th, 2007 14:00

Which guide are you looking at? If you are looking at the latest guide then the clients will still connect to the Exchange server name. No alias needed. The latest guide has you rename the servers and do a disaster recovery install of Exchange on the target after it has been renamed to the production Exchange name. After everything is done, the two physical servers will have different names from that of the Exchange server in ESM and AD.

So the clients will still point to the server name that shows up when you go into your Exchange Server Manager. The DNS and Netbios resultion will be controlled by the module.

As far as what the clients will see, if it is Outlook XP(2002) you will have to restart Outlook. Outlook 2003 will usually tell you something like (Your Exchange administrator has made changes that require you to restart Outlook). I have seen Outlook 2003 just reconnect without any messages though.

The biggest thing to look for on your failover with this is the DNS cache on any front end servers and the client machines. Some sites it can take 15 minutes for the cache to time out. Asking all users to reboot would take care of all your issues. Otherwise you can just wait for the cache to timout or run ipconfig /flushdns.
No Events found!

Top