asked on

Exchange 2013 Autodiscover and redundancy and an IP question

I am configuring Exchange 2013 but I have not been able to find documentation on the Autodiscover DNS record. If autodiscover points to the mailserver, and that mailserver goes down, how does it know where to fail over to? Do I need a second Autodiscover record? in the same vein, how many public IP addresses do I need for Exchange 2013? Is it one to handle Autodiscover requests and one for everything else? or can everything be done over port 443 on a single IP address? is there a way to split this out, asking because it appears that our old 2010 infrastructure achieved this (and I don't like it!).
Thanks!

SOLUTION

Ivan

membership

This solution is only available to members.

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

Start Free Trial

it_medcomp

ASKER

That's what I thought, but here is the part I don't understand. Let's say Autodiscover resolves to 123.231.222.111, which happens to be a public IP that the firewall redirects to an internal address 10.10.32.1. 10.10.32.1 goes down. That means effectively that the IP Autodiscover resolves to, 123.231.222.111, becomes unresponsive. How do clients know about the other exchange servers at 10.10.32.2 and 10.10.33.1? This assumes the second site is behind the same firewall as the primary site.

I guess it maps out like this:
Server Public Private
EXCHANGESITE1-1 123.231.222.111:443 10.10.32.1
EXCHANGESITE1-2 10.10.32.2
EXCHANGESITE2-1 10.10.33.1

So the Autodiscover record points mail.contoso.com to 123.231.222.111- With DNSRR as our 'load balancing' strategy, what tells Autodiscover to look at another Exchange server- I don't believe our firewall cares about our internal DNS, and the address is NATted to the IP of the first server.

Simon Butler (Sembee)

DNS does not have any redundancy. It has no knowledge that a host is available or not.
If you want to have redundancy then you will need a load balancer to connect the clients to, which will be host aware.

You can give each Exchange server an external IP address and add them to the DNS record. However that will not give you redundancy either. It will be pot luck If a user hits a server that is live or not.

it_medcomp

ASKER

OK, so on a similar vein, if I have three sites, A, B and C, and I sit down at A and configure my Outlook client, and sit there for two months, then I go to site B and sit there for a week and then spend a week at C, does Exchange use the AD Site to determine which mailserver I connect to, or is this determined some other way? If so, how do I configure site affinity for Site D, which should use site B's server, assuming the latency is lowest between those two sites than of D to C or of D to A?

ASKER CERTIFIED SOLUTION

Simon Butler (Sembee)

membership

This solution is only available to members.

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

Start Free Trial

it_medcomp

ASKER

There were two questions here and they were both answered. Thanks to all who contributed!