Unsolved
This post is more than 5 years old
2 Intern
•
143 Posts
0
5519
September 13th, 2012 12:00
Avamar 6.1 Exchange 2010 DAG configuration
We are running Avamar 6.1.0-402 and would like to setup federated backups of our Exchange 2010 DAG. We have three production mailbox role servers that house live mail databases; all servers are Windows 2008 R2 Enterprise. Exchange 2010 has SP1 and update rollup 6 installed. Each database has an active copy on one mailbox server and a passive copy on another server at this site.
Looking at page 71 of the EMC Avamar 6.1 Exchange VSS User Guide, it looks like Avamar will ask for "DAG group for Exchange DAG client", "Exchange DAG client network name", "Exchange DAG client IP address" and a netmask for that IP. What is the purpose of each of these? I'm particularly confused by the IP address it asks for and the Exchange DAG network name. Are those simply the name of the DAG and its IP? The documentation glosses over these points and the terms it uses are crazy confusing! I've setup a DAG before in CommVault and Backup Exec and this is by far the most perplexing. Please help.



Amol
2 Intern
•
207 Posts
0
September 13th, 2012 12:00
In order to configure DAG client for Avamar, you need a spare IP address which is NOT used on the network.
In the background it will create failover Cluster Resource on all the Exchange Server and add "Avamar Agent" under "Services and Application" . Avamar agent will use the IP address which was provided during the setup
Once the cluster resource is created it will also create a new Client (DAG) on Avamar
All the backup request for the DAG will be send to this Client new Client(DAG), this client will indeed determine who owns the passive copy,then it will forward the backup request to desired Exchange Server and backup will start.
All the backups will be tagged to the newly created Client.
Adding a screenshot for your reference
fdxpilot
2 Intern
•
143 Posts
0
September 13th, 2012 14:00
Thanks for the screenshots, Amol. They clearly explain what is really going on behind the scenes here and would be good to add to the documentation, which is currently quite vague.
But, wow, that is totally different than how any other backup product of which I'm aware handles backing up Exchange databases in a DAG configuration! No product of which I know requires modifying the DAG failover cluster. Instead, products I've used before rely on ADDS and Exchange PowerShell commands to discover which servers host copies of a database, chose which server to run the backup against, and then connect to the chosen server to execute the database backup commands via VSS. Do you have any idea why Avamar selected this more complex approach? It also appears this method may not be supported by Microsoft. I found this quote from Microsoft TechNet:
Amol
2 Intern
•
207 Posts
2
September 17th, 2012 10:00
You are correct, using an existing failover resource is not supported by microsoft and therefore we create NEW failover group "EMC backup Agent for xxx" by default. We set up Avamar to attach to an AvAgent in this new group, and use AD and PowerShell to find the correct physical nodes, then use Avamar’s InterAgent CTL (federated) to attach to the correct physical server(s) and do backups, and snapview it all together from our group
We have worked with Microsoft to ensure that this is supported; that is why we create a separate failover group rather than inserting our AvAgent into the same group that is created for the DAG (which is what legacy Exchange plug-in used to do).
scoopat
3 Posts
0
April 5th, 2013 12:00
Thank you for this information Amol. As the Original Poster mentioned, the EMC documentation on this is not so clear.
So you need a spare IP address. Does it need forward and reverse DNS entries to work??