Exchange 2010 Redundancy

BoomAM

BoomAM

BoomAM

Soldato
Joined
5 Jul 2003
Posts
16,206
Location
Atlanta, USA
Hi,
I'm currently considering a hypothetical project to allow multiple site exchange connectivity whilst providing protection against site failures.

My current idea is as follows:
All sites to have CAS, MB & HT, in addition to a DAG.
The reasoning being that it will allow local users to the site quick and seemless access to their mail, with the DAG groups and inter-site links providing a level of failover should a mailbox, CAS or HT fail.
& an edge server & CAS sat in a hosted datacenter, to allow outgoing mail between sites and a central CAS server for remote workers & devices that can access MB & DAG servers at any site to get users mail.

Does anyone have any comments/suggestions on that model?

Thanks in advance all. :)
 
Makes sense but obviosuly it is an expensive and more complicated option, if email really is that important to the business then it maybe the correct solution.
 
If you're going with a single box at each site for all those roles, that's fine.
You will need Win2k8 R2 Enterprise or better for DAGs.

One issue with that is that if the box goes down, you lose the local CAS role and as such, will suffer an outage. To fix this, you'll need a CAS Array.
Putting a standalone instance of Exchange 2010 Standard running just the CAS role so it can repoint it to the other servers will fix this. One VM instance and one CAS array at each AD site running a CAS role that you want failover to occur at.
You need to set up the CAS Array before you make any Mailbox Databases on that site, as there is an attribute field in the Mailbox DB which tells it where requests will be coming from. (Although, from a mailbox portability point of view, maybe this changes if you dismount + remount the DB elsewhere? Anyway, that's best practice!)

The use of CAS Arrays also makes it easier incase you need to expand the number of roles / machines carrying out the roles at any of the sites.

I would recommend reading up on the Failover, High Availability and Array features of Exchange 2010 for getting a better idea of how it all slots together.
 
If you do go down cas array remember you need load balancing to go along with it either hardware load balancer, a virtual load balancer or software load balancing (WNLB is not very good and can't be used on same server as a DAG)
 
The best solution is usually the simplest in my experience.

A second, lower priority MX record pointing to a mail server that

A. periodically tries to deliver mail to the primary mail server.
B. provides users access to their mail during the interim period that the primary mail server is down.

This is how I've done it in the past, and it can be done for peanuts with a VPS and a little knowledge of Linux and one of the commonly available MTAs.

You can even have it authenticate against your AD, should you wish.
 
Last edited:
The MX record route is an interesting one, and is perhaps something that can be used in addition to the option I'm considering.
 
or an off host solution? we use mimecast for spam filtering. but they can be used as a webmail interface for dr in the event your main server / office / site goes down. as they are the mx record there is no updating only telling clients to go "here for webmail" - they also have an "outlook connector" but tbh I havent tried it.
 
You must log in or register to reply here.
Back
Top Bottom