Regarding the second question, I have noticed it and wondered the same thing. The only reason I can think of is the performance issue, two specifications (one for database and the second for the transaction logs) can work in parallel which can help with the busy Exchange server...otherwise it should work just fine if they are both handled by the same specification. Of course thet doesn't explain why they need to be on the separate *partitions*.
For the failover to work correctly I would expect the failover box to be a member of the domain so I would expect that you would need to join the existing domain.
I believe that the only reason you would split the logs and EDB files would be for performance; it should work on the same partition but if you have a heavily utilised Exchange box it may have issues.
You would be sensible in this configuration to employ a failover agent to prevent a split brain configuration - if there is an issue with the connection but not the server it is possible that the failover box will take over from the production box and both will continue running - which is sure to lead to problems. A failover agent is another box that both servers can see which will guard against this problem occuring.
Answering your second question, we have Replistor installed on 15 servers at our main office and 15 servers at our disaster recovery site. To be able to replicate the data, we have found that joining the existing domain at our main office makes the replication process a breeze! Good luck.
I notice this topic has been here for a bit but I have a specification running between servers in different domains. But if it is feasible for you to join both servers into the same domain I would go that route
If you are doing a failover of Exchange using a RepliStor Alias, one of the steps to setting up the environment is to join the DR Exchange box to the same Organization as the Production box. Therefore, it will need to be in the same domain. The exch2k.dll will update AD with specific data for the failover to the other node.
For this to work, the paths of the Specification must also be identical. There are entires in AD that also point to these paths and if they are different the Alias failover will not work correctly.
tribicic
157 Posts
0
January 26th, 2006 10:00
bmartinik-axnd.
8 Posts
0
January 26th, 2006 10:00
DavidHampson
2 Intern
•
1.1K Posts
0
January 30th, 2006 03:00
I believe that the only reason you would split the logs and EDB files would be for performance; it should work on the same partition but if you have a heavily utilised Exchange box it may have issues.
You would be sensible in this configuration to employ a failover agent to prevent a split brain configuration - if there is an issue with the connection but not the server it is possible that the failover box will take over from the production box and both will continue running - which is sure to lead to problems. A failover agent is another box that both servers can see which will guard against this problem occuring.
slemburg
6 Posts
0
February 27th, 2006 07:00
bmartinik-axnd.
8 Posts
0
February 27th, 2006 16:00
bfv5nl0
1 Message
0
April 4th, 2006 11:00
dramjass
151 Posts
0
June 9th, 2006 12:00
For this to work, the paths of the Specification must also be identical. There are entires in AD that also point to these paths and if they are different the Alias failover will not work correctly.