Deploying Disaster Recovery for Exchange 2010 – Part3
Dec 21, 2011 Microsoft Exchange Server
Deploying Disaster Recovery for Exchange 2010 – Part3
-Authored by Achinta Chatterjee (link to Linkedin Profile)
Introduction
In this document we will go through the steps to execute a DR drill and an actual DR scenario. We have already detailed in the Part1 and 2 on Client connectivity and Mail Routing for Primary and DR site.
DR Drill
The DR Drill exercise can be divided into three sections:-
i. Preparation Steps for DR Drill
ii. DR Drill Execution Steps
iii. DR Drill Testing
i. Preparation Steps for DR Drill
1. Create a Mailbox database for DR drill purposes.
Once this database is created make it a part of DAG.
Make sure the RPCclientaccessserver for this database is the DR CAS Server. For eg, you can use the following command to set the RPCClientAccessServer for DRDB01 database to the DR CAS HUB Server.
Set-MailboxDatabase -identity DRDB01 – RpcClientAccessServer temexgcasr01.thdm.local
Temexgcar01.thdm.local is the DR CAS HUB Server.
2. Activate the database DRDB01 to DR site Mailbox server and swing it back to the Primary site. This is to test whether the database swing is working fine between the 2 sites.
3. Identify those production users who will participate on the DR drill
4. List domain names the DR drill users will be sending emails to the Internet.
We need to find this as we need the drill users to send emails to the Internet from the databases hosted on the DR servers and through the DR Internet Mail relays. As this is a separate exercise to test DR readiness we won’t affect the production users send and receiving emails.
In our DR drill we will use gmx.com as the domain name to which DR drill users will send outgoing emails. To accomplish we will need to create a Send connector entry on the Send Connector “DR SMTP Outgoing” in Part 2 of this article. Click Add in the Address Space tab and put the address gmx.com with a Cost of 1. Note the purpose of the “Cost” field here. We already have another send connector with * with a cost of 1 for the Primary Site. All Outgoing Internet emails go through that connector. But when you specify a domain name (for eg xxx.com) it takes priority over * and hence any emails for gmx.com will route through the DR site Internet Mail relays.
ii. DR Drill Execution Steps
1. Notify DR Drill users to exit their Outlook Clients and not to access emails.
2. Move all DR Drill participating users (identified in step 3) mailboxes from their Primary databases to the DR database DRDB01.
3. Activate the database created for DR drill purposes DRDB01 to be mounted on the DR Mailbox server.
4. All DR drill user mailboxes will failover to the DR site.
5. Notify the DR drill users to open their Outlook clients. The Outlook client when opens for the first time after the database has been moved to the DR site will take a few more seconds. The reason being now the Outlook client configuration will be pointing to the DR CAS Server (temexgcasr01.thdm.local) and Outlook anywhere URL will be dr.myemail.com.sg. Figures are show below.
iii. DR Drill Testing
Once the users have connected to the DR site they need to test the following items.
1. Outlook Connectivity – as shown in the previous section.
2. From Outlook client test sending and receiving emails between DR drill participating Exchange users.
3. Send Outgoing emails from Outlook client to gmx.com. The administrators need to check from the server logs whether these emails are routing through the DR Internet Mail Relay servers and confirm.
4. Reply emails from gmx.com to see whether they are being received back to the sender. Note this routing will be through the Primary site.
5. Administrators need to inject emails manually from the DR Incoming Internet Mail Relay servers to the DR drill users. This is done as the actual MX is still pointing to the Primary site. Confirm that these emails are received by the DR Drill users. This will confirm that the Incoming internet emails are being delivered properly from the DR Incoming Internet Mail Relay servers, and should be functioning during the actual DR.
6. Open DR Site OWA URL http://dr.myemail.com.sgand confirm the user can login and open mailbox.
7. Test step 2,3,4,5 again from OWA just to reconfirm your testing.
Disaster Recovery (actual scenario)
During actual DR we need to execute the following steps.
1. Outlook and OWA client connectivity will be lost all users. Ask all user s to exit their Outlook and OWA clients
2. When the Primary site is down all Mailbox databases will auto-failover to the DR mailbox server. The reason being the Activation preference on these databases is set to the DR mailbox server as the last resort. Confirm that all databases have been mounted on the DR mailbox server.
3. Take the list of all mailbox databases you are hosting and run the command below from the EMS for each database on the DR mailbox server. Replace the “databasename” with appropriate databases.
Set-MailboxDatabase -identity “databasename” – RpcClientAccessServer temexgcasr01.thdm.local
4. Once the above is completed open EMC and browse to Organization Configuration >Hub Transport> Send Connectors . Open the Send Connector already created “DR SMTP Outgoing” Properties. Change the Cost of the connector from 100 to 1.
5. Once the above is completed open EMC and browse to Organization Configuration >Hub Transport> Send Connectors. Open the Send Connector for Primary Site Mail routing “Outgoing” Properties. Change the Cost of the connector from 1 to 100. The above 2 steps will make sure that the Outgoing Internet emails will be routed through the DR Internet Mail Relay hosts.
6. Once the Primary Site is down the MX preference for the hosted domain names will make sure that incoming Internet emails get routed automatically to the DR Incoming Internet Mail server. As shown in the figure below postmantp1b is the DR Incoming Internet Mail server with the highest MX preference of 40.
7. All emails will start queuing on the DR Incoming Internet Mail server. Once the Exchange DR CAS HUB server is available the emails will get routed out to these servers and eventually to the respective user mailboxes hosted at the DR mailbox servers. To ensure all mails is getting routing properly check the incoming logs of the DR Incoming Internet Mail server and the Queue viewer on the DR Exchange CAS HUB servers, as shown below.
8. Once the above has been accomplised the administrators should do a few rounds of DR testing. Verfiy Outlook client connectivity by opening some Outlook clients and sending and receiving emails between themselves and the internet. Verfiy OWA URL dr.myemail.com.sg is accessible and can open mailbox databases.
9. Once all tetsing is confirmed notify the users to open their Outlook Clients or OWA (using the DR URL dr.myemai.com.sg). Business should be back to normal as far as end users are concerned.
Conclusion
This marks the end of the 3 part article on Deploying Disaster Recovery for Exchange 2010. We have tried to detail Exchange 2010 DR configurations, Client Connectivity, Mail Routing, DR Drill and Actual DR. You can customize you DR Startegy w.r.t your own environment with the above recommendations.
No Comments