Home | Contact Us | FAQ | Search & Site Map | Link to Us
Sign In | Join | Other 45 Sites in Network
Home
Discussion GroupsWindows Server 2003Windows 2000Windows NTSmall Business ServerVirtual ServerExchange ServerIISHost Integration ServerISA ServerSMSWSUSMOMWindows Media ServerSecurityCertification
Related Topics
SQL ServerMS WindowsMS OfficePC HardwareMore Topics ...

Windows Server Forum / Exchange Server / Design / September 2008

Tip: Looking for answers? Try searching our database.

Exchange 2007 CCR Design

Thread view: 
Enable EMail Alerts  Start New Thread
Thread rating: 
Ian - 16 Sep 2008 08:26 GMT
Hi,

I need some help with our exchange 2007 CCR design.  We have 2 production
sites that act as DR sites for each other.  One site is in New York, the
other is in London and are connected via a 45MB DS3.

Initially, we've setup the sites as follows for testing.  So as to reduce
the number of servers involved we've use a single site for London and New
York exchange implementations.

In New York we have

Primary New York Exchange 2007 CCR Node
Passive London Exchange 2007 CCR Node
2 x HUB/CAS servers with the CAS server using network load balancing
1 Domain controller



In London we have

Primary London Exchange 2007 CCR Node
Passive New York Exchange 2007 CCR Node
2 x HUB/CAS servers with the CAS server using network load balancing
1 Domain controller

We have a spanned vlan between the sites but we have allocated a different
vlan for New York devices and London devices although both subnets are
included in the same Active directory site.

Now, our problem at this point is that our domain controllers can not
replicate because they are in the same site but are located in different
subnets and therefore cannot replicate via RPC.  If we change the replication
to IP based it works for about 1/2hr but then it automatically changes back
to RPC - we cannot make it stick.

The other problem is that we often find the exchange CCR active node in New
York will actually be using the London based domain controller which is not
too bad in testing but once we add this to production with a larger number of
users we want to make sure that the exchange nodes query the AD server in the
physical site.

We realise that we could solve this by having seperate sites but our goal
was to reduce the number of servers involved so that the active domain
controller in one site would also be the backup domain controller for the
passive CCR node in failover and also so that the hub servers in the other
site could provide redundancy in the event of a full site failure with no
re-configuration required.

For example, we could have this configuration.

AD Site LON-NY would include (all on the same subnet via the stretched vlan)

In London:

Active Exchange mailbox CCR Node
Hub/Cas server
Domain Controller

In New York:

Standy Exchange mailbox CCR Node
Hub/Cas
Domain Controller

---------------------------------------------

AD Site NY-LON would include (all on the same subnet via the stretched vlan
- different subnet to the LON-NY site obviously)

In New York:

Active Exchange mailbox CCR Node
Hub/Cas server
Domain Controller

In London:

Standy Exchange mailbox CCR Node
Hub/Cas
Domain Controller

Now the problem with this solution is we would end up with 2 extra domain
controllers and we don't have CAS redundancy like we do in our initial test
where we have a CAS NLB in the same site.

Any advice greatly appreciated - sorry for the long post.

Thanks,

Ian
Ed Crowley [MVP] - 17 Sep 2008 04:55 GMT
What kind of quorum are you using and where is it located?

Do you really want your Exchange server automatically failing over across
the ocean?

Are you absolutely sure you don't want to use SCR for this?
Signature

Ed Crowley MVP
"There are seldom good technological solutions to behavioral problems."
.

> Hi,
>
[quoted text clipped - 94 lines]
>
> Ian
Ian - 17 Sep 2008 09:02 GMT
I know this is pushing the limits of CCR's high availability but do want to
give this a shot and so far our testing has shown that it works well.  We
don't just want this as a DR solution - we want to use it for when we apply
updates and perform maintenance in the production site etc so will failover
regularly.  

If this doesn't work out we will switch to our Recoverpoint replication
model where we replicate the entire exchange virtual machine and power on an
identical copy in the remote site for failover - this requires a slightly
more manual process than CCR but not as much as SCR requires.

We are running Windows 2003 x64 R2 SP2 and are using a Majority Node Set
Quorum with the file share witness located on one of the local HUB servers.

> What kind of quorum are you using and where is it located?
>
[quoted text clipped - 100 lines]
> >
> > Ian
 
Sign In
Join
My Latest Posts
My Monitored Threads
My Blog
My Photo Gallery
My Profile
My Homepage

Start New Thread
Enable EMail Alerts
Rate this Thread



©2009 Advenet LLC   Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.