Exchange 2010 DAG Backup

[johntnavi]

Hello,

We are going to be migrating to Exchange 2010 in the coming weeks. We will be using a physical server as the Active node and a VM as the DAG Passive. Can I just backup the Passive VM and retain brick level recovery?

[andersonts]

Yes you can...you will want to make sure to enable application aware processing but also go into the advanced settings and make sure you don't truncate the logs...let Exchange handle that for you.

[chrisdearden]

Even better , by backing up the VM node , you'll be able to use our Exchange Explorer feature for fast and easy item level restores from Exchange 2010.

[andersonts]

Absolutely...this is actually a very good template for others to use as this comes up quite a bit. If you haven't seen Veeam Explorer in action check it out here: http://www.veeam.com/microsoft-exchange-recovery.html

I have used it on a 1TB mail store and it was ready to go in about 3 minutes. Good stuff!

[foggy]

A couple of existing topics on the subject, could be helpful:
Exchange 2010 DAG backup ?
Backing up Exchange 2010 - How to truncate logs?

[andersonts]

I wanted to clarify that if you are only backing up the passive node you should leave log truncation enabled. The passive node will then replicate the log truncation request to the other nodes.

[trackstar]

Hello,


Currently we have one Exchange 2010 mailbox server (role) and two Exchange 2010 CAS/HUB servers. We are in testing phrase but will eventually have more than one mailbox server.

I have tested been testing the Veeam Backup and Replication on that one mailbox server. We will like to add a few more mailbox servers and go with DAG topology. How would I plan my backup? Would Veeam backup just the active databases? Also, would the implementations affect any licensing issues?

Currently we have the Enterprise version of Veeam

TIA,
tntrac

[Bunce]

I wanted to clarify that if you are only backing up the passive node you should leave log truncation enabled. The passive node will then replicate the log truncation request to the other nodes.

It should be enabled irrespective of whether you are backing up a passive/active or hybrid node.

Truncation will be replicated across the DAG irrespective (as long as conditions are satisfied for all).

[mattk0]

We are implementing DAGs in our Exchange 2010 environment and had some confusion about how VEEAM works with the log truncation. Email1 and Email3 (site A) are backed up at 5pm and Email2/4 (Site B) are backed up at 10pm, all transaction logs are truncated by VEEAM. If VEEAM truncates the logs of email1db and 2db (passive) on email1 at 5pm, wouldn't that mean the passive database logs no longer match the logs on the active server (email2) which won't be backed up and truncated until 10pm?

If so - is this acceptable? I'm concerned if there is an outage between 5-10pm, the transaction logs will look different on the email servers because VEEAM already truncated one but not the other?

Our scenario is this:
SITEA SITEB
Email1 | Email2
Email1db (active) | Email1db (passive)
Email2db (passive) | Email2db (active)

Email3 | Email4
Email3db (active) | Email3db (passive)
Email4db (passive) | Email4db (active)

[foggy]

Matt, please review the topic above, should answer your questions. Feel free to ask for any additional clarification, if required.

[denonsix]

We're converting a 2 node physical Exchange 2010 DAG into a new config where at least one of the nodes is a VM (ESXi) so it can be backed up with Veeam.

Is there any reason to make both of the nodes VMs (from a backup perspective)?
I am assuming the best plan is to make the VM node the passive copy and just back up that in Veeam, the capturing all necessary backup data to restore Exchange data.
The active copy would not be backed up.

Keeping the active node physical saves on the licensing and small VM overhead and deduplication store space (of not backing it up). I also understand backing up both can lead to occasional failovers due to snapshot processing. However, are there any advantages to making both virtual and backing up both with Veeam worth noting?

[foggy]

Please review the posts and the threads mentioned above for considerations regarding backing up Exchange DAG. Feel free to ask for additional clarification, if required.

[Ultra_Cool]

Hello,

We have 2 Exchange servers and they are DAG members. If one of them crash how will I restore it and not lose replication (We have mirroring on two Exchange servers). Shall I restore whole VM and let Veeam handle everything or is there something I should think before I restore it?

[veremin]

Your post has been merged into existing discussion regarding DAG backup. Please, see answers provided above and ask for additional clarification in case you need it. Thanks.

[Ultra_Cool]

Hello,

Our Exchange servers are both VM's and we want to protect only active databases that means only our primary exchange. Now what will happen with replication after restore if that server crashs? Shall Veeam fix everything or I need to do additional steps?

[foggy]

As it is mentioned in the topic referred to above, you would be able to restore the failed server with the Exchange recovery option.

[garethf]

Hi team

Just a quick query, we have a large-ish Exchange 2013 environment and are looking at some real world best practices/guidance around backing up with Veeam.

Our DAGs are 10-12 node stretched multi-site clusters, each node has 20 x ~400-500 GB mailbox databases on them (13 or so per node active). At the moment physical but we are looking at virtualizing with Hyper-V to make use of Veeam.

We have other Exchange 2013 2-node DAGs that Veeam is backing up fine, but when it comes to larger clusters some questions come about. A lot of the guidance seems geared for smaller environments.

Would the best option in these larger environments generally be to have passive nodes (e.g. at secondary sites) and back up exclusively off those? And not back up the active servers. Our veeam reps are not using deduped backend storage. If that's the case we could leave our active servers physical which could work out well.

I've read the Veeam guidance, but interested in what are other people doing in the real world