Issue failing over, clients can't connect! (Full Version)

All Forums >> [Microsoft Exchange 2010] >> High Availability



Message


dclipp -> Issue failing over, clients can't connect! (22.Aug.2011 12:01:33 PM)

Our setup: We have 3 servers in our Colo, there is a CAS, HT, and Mailbox server. In our corporate office we have a server that is configured as all three roles. The mailboxes as setup as a DAG and all replication is functioning as expected. The HT in the Colo is the witness server. The MB server in the Colo is the active server. This weekend we tested Exchange failover functionality in the event we lost the Colo. The active mailbox was transferred as expected and OWA worked as well. The problem is that no Outlook clients would connect to the server in the corporate office. What am I missing here? Does anyone have any ideas? Thank you for your help!




de.blackman -> RE: Issue failing over, clients can't connect! (23.Aug.2011 9:28:59 AM)

Unless I have misread\misunderstood the scenario, your Outlook clients are pointing to the name of the CAS server in the colo and when you did a failover simulating a loss of the colo site, that CAS server is no longer available (you can confirm by trying to ping the CAS server name when you failed over to the DR site). What I would recommend is to change the IP address of the CAS server name in your internal DNS to point to the IP address of the CAS server in the DR (corporate office) site.

How did the OWA clients connect? What URL did they use (same as they had always used)?




dclipp -> RE: Issue failing over, clients can't connect! (23.Aug.2011 9:58:47 AM)

You understand correct. We simply failed the DB's to the DR location and powered off the server in the Colo. With that done we changed internal DNS to point mail.domain.local to point to the DR server and OWA worked as expected. When opening Outlook client's we attempted to connect them by point Outlook at the DR server. They were unable to connect... Please let me know if there are any additional questions that can help to clarify the situation. Thank you for your help!




travis.sheldon -> RE: Issue failing over, clients can't connect! (23.Aug.2011 10:08:31 AM)

Did you confirm that Outlook Anywhere (if used) was setup in the failover datacenter? Is the domain controller on Exchange in the primary site?




dclipp -> RE: Issue failing over, clients can't connect! (23.Aug.2011 10:13:10 AM)

We are not using Outlook Anywhere. We are simply troubleshooting internal connectivity at this point. Each site has 2 Domain Controllers. Exchange is not installed on any Domain Controllers.




travis.sheldon -> RE: Issue failing over, clients can't connect! (23.Aug.2011 10:19:42 AM)

Im wondering if it's the RPCClientAccessServer setting on the Mailbox Databases -- run this to make sure it's pointing to the same record you adjusted to the failover server

Get-MailboxDatabase | fl rpc*




dclipp -> RE: Issue failing over, clients can't connect! (23.Aug.2011 10:44:27 AM)

That command is currently returning the name of the CAS server in the Colo as we would expect. When we failover and shutdown all of the Colo servers do we need to manually change this to point to the CAS in our DR site? If so, how?




travis.sheldon -> RE: Issue failing over, clients can't connect! (23.Aug.2011 11:11:40 AM)

You shouldn't have to if the DNS was adjusted..

For instance, if you use cas.domain.com and it points to the primary site CAS server and you experience an outage, then you'd adjust the cas record to point to the secondary site CAS server..this way you don't have to adjust the RPC Client Access server values, etc.

Just to make sure, you don't have any entries in the hosts file, specifically on the passive mailbox server, correct?

I believe that Outlook is either still trying to send MAPI requests to the downed CAS server in the primary site or trying to send the authentication request to the DCs in the primary site.. unfortunately the only way I know to test it is by reactivating the fail over and either running Outlook from command prompt "outlook /rpcdiag" or running testexchangeconnectivity.com




dclipp -> RE: Issue failing over, clients can't connect! (23.Aug.2011 11:20:22 AM)

So, this is probably an error on my part. The client's are pointed to servername.domain.local instead of CAS.servername.com. Do you think that if I created a CNAME record for servername.domain.local and pointed it to backupCAS.domain.local the clients would be able to connect?




travis.sheldon -> RE: Issue failing over, clients can't connect! (23.Aug.2011 11:39:56 AM)

Technically they should be able to connect...but in all honesty I'd look at setting up a CAS array...even if you don't physically setup the load balancing between CAS severs, you can use the ClientAccessArray in active directory to set a custom FQDN for Outlook/Exchange services..

For instance, if you run new-clientaccessarray and set the FQDN to outlook.domain.com, then whenever you have a fail over you'll modify the DNS record to point to the backup server..all clients either external or internal would connect to outlook.domain.com (you'd have to create a DNS record internally/externally for it) which allows you to freely move the destination if needed.

If you do what you described above and modify the CNAME, it could have unexpected results in the primary site when the servers try to communicate with the actual mbox1 server (which would then be CNAMED to mbox2)




dclipp -> RE: Issue failing over, clients can't connect! (23.Aug.2011 1:30:01 PM)

Do you have a source for good documentation on creating a CAS Array. It sounds like the right way to go.




travis.sheldon -> RE: Issue failing over, clients can't connect! (23.Aug.2011 2:01:44 PM)

http://howtoexchange.wordpress.com/2009/12/16/configuring-client-access-array-for-exchange-2010-walkthrough/

Skip step 2




Page: [1]