Restore Exchange Mailbox Items from Storage Snapshot

  • 29 September 2023
  • 5 comments
  • 112 views

Userlevel 7
Badge +8

Using Veeam to orchestrate storage snapshots is a fast, reliable way to have frequent restore points and quick restores.  Unfortunately, unlike a VMware backup, storage snapshots on databases are not application aware.  Restoring application items from a storage snapshot will often give you a blank explorer window.

 

The first thing you will want to do is restore from the storage snapshot, and select a Host to mount the SAN volume on

 

 

With the explorer up it will look empty, but the volumes will be mounted to your PC.

Click Add Store, or right click All Stores and Add a Mailbox Database

 

 

 

Next, we will point the explorer to the specific Exchange mailbox database.

 

 

 

If you go to C:\VeeamFLR\*VMname* you can see Volumes listed as folders there that are the mount points to the disks. This should mirror the folder structure on your Exchange server.

 

In this specific situation here is what it looked like.  Depending on your disk configuration and folder structure the volume, VM name, and folder will be specific to your environment. (As your Email Admin if you need assistance)

 

Here is the location of the mailbox DB I was looking for.

C:\VeeamFLR\*VMNAME*\Volume2\DBs\*MAILBOXDB*

 

You then have to map the log folder which in my environment is located on a different volume.

C:\VeeamFLR\*VMNAME*\Volume7\logs\*LogFolder*

 

 

Finally, depending on the state of the database during the snapshot, you may need to map ESE.dll which should be on the Exchange server as well.  I have had it happen where this is required, and where it isn’t required.  

https://helpcenter.veeam.com/docs/backup/explorers/vex_ese_config.html?ver=120

 

 

This is the location on the Exchange server.

C:\Windows\Installer\$PatchCache$\Managed\442D7VJ38B9ASADF0962A6BFFFFFFF1F\15.1.2507

 

I found that maps to the following location on my PC while doing the restore

C:\VeeamFLR\*VMNAME*\Volume1\Windows\Installer\$PatchCache$\Managed\442D7VJ38B9ASADF0962A6BFFFFFFF1F\15.1.2507

 

Veeam will now show the Explorer for Microsoft Exchange just like a regular backup and allow you to restore from a storage snapshot.

 

 

 


5 comments

Userlevel 7
Badge +21

Love storage snapshots and how fast they are compared to standard restores.  Working to get Hitachi working in our environments.

Great post by the way.

Userlevel 7
Badge +6

Wait...Exchange on-premise?  Who does that anymore?

(I kid...I have two clients that are still on-premise despite my best efforts.  And one (former client) that was/is running Exchange 2010 - that’s gotta be hacked by now, right?

Seriously though...helpful info here...thanks for posting this!

Userlevel 7
Badge +19

Nice succinct, intuitive post on restoring from Storage Snaps Scott. I think I had to recover a VM that way once (I think because I deleted the b/u job as it was no longer relevant...so we thought). Appreciate you sharing.

Userlevel 7
Badge +8

Wait...Exchange on-premise?  Who does that anymore?

(I kid...I have two clients that are still on-premise despite my best efforts.  And one (former client) that was/is running Exchange 2010 - that’s gotta be hacked by now, right?

Seriously though...helpful info here...thanks for posting this!

Fully on-prem here. we are working on it.

Unfortunately, our email addresses are on another domain managed by another government entity, so it’s not an easy transition. Hoping to be 365 on our own domain within the year. 

Userlevel 7
Badge +8

Nice succinct, intuitive post on restoring from Storage Snaps Scott. I think I had to recover a VM that way once (I think because I deleted the b/u job as it was no longer relevant...so we thought). Appreciate you sharing.

 

I had a request to restore an email from someone at the top of the food chain. They received the email minutes after the backup ran, but I had snapshots enabled on the volume for this particular reason. I didn’t know why the explorer was empty but after sorting it out recorded this information to speed things up next time.  I tried it again and it wasn’t required, so I think it’s hit or miss if the database needs a rebuild. 

Comment