Best Practice for Exchange Server 2016 Backup

[salimhurjuk]

Can you please assist me to understand Best Practice to take Backup for Exchange Server 2016?

What things should be consider during the Backup for Exchange Server 2016?

[Andreas Neufert]

Please read through this article and come back with questions:
https://andyandthevms.com/exchange-dag- ... plication/

[MPECSInc]

First, if the Exchange Server is virtualized do not use Snapshots/CheckPoints with the exception of running updates. But, one should _always_ set the SMTP 25/587 to DISABLED at the edge prior to starting the updates.

If using a Snapshot/CheckPoint prior make sure to delete it after the updates have completed and re-ENABLE the port forward for SMTP.

NOTE: It's a good idea to have a SmartHost and continuity service in front of Exchange Server. SmartHost preserves the server's outbound reputation but also protects the Exchange Server from SMTP Auth attacks as the Edge would be set up to only port forward from the continuity provider's servers/subnets.

Exchange Server should be set up according to best practices to allow for maximum recoverability:
* 150GB OS
* 250GB Exchange Server install
* 500GB Exchange Logs
* 500GB Exchange Databases (Recommended maximum size is 250GB per database)

If there's a need, have several .VHDX/.VMDK files attached to the VM to satisfy the above requirements.

Backup: Make sure to use a frequency schedule that meets recovery point objectives.

Make sure to enable in-guest acquiescence to trigger log consolidation by Exchange Server. If this setting is not configured, then Exchange Server does not know it is being backed up. Thus, it won't consolidate the logs.
One could enable circular logging as an alternative.

Also keep in mind the age of the hardware and the storage I/O capabilities. If there's an I/O struggle during backups make sure to tune Veeam to storage latency so as to not impact the Exchange Server's ability to receive and send mail.

The above is off the top. If you have any other questions please drop them here and we'll see about what needs to be said. ;0)

[salimhurjuk]

Which logs from Exchange Server do you consider taking the Backup?

[dloseke]

Exchange Server should be set up according to best practices to allow for maximum recoverability:
* 150GB OS
* 250GB Exchange Server install
* 500GB Exchange Logs
* 500GB Exchange Databases (Recommended maximum size is 250GB per database)

If there's a need, have several .VHDX/.VMDK files attached to the VM to satisfy the above requirements.

This sizing is...large...possibly. I mean, some of it makes sense, but for small environments with say, 50 users or something like that may require much small disks. For instance, I have one client running on a 50GB log drive. Assuming nightly (or more frequent) backups, and based on the amount of mail flowing through the server, this may be more than sufficient. I'm not sure what the 250GB disk is needed for on the install drive unless you're trying to account for the transport queue, but that can also be relocated. Exchange database drives are as needed based on the database sizes. My point is you're speaking in absolutes, and there's often not a one-size-fits-all sort of solution for this, so without knowing more about what the users requirements are, it's hard to make generalized statements here.

For me personally, but depending again on the size requirements of the users and expected mail flow, but this is where I start unless I know I'm going to need more....
C: 100GB (OS)
D: 100GB (Exchange Server Install)
M: 250GB (Mailbox Databases)
L: 50GB (Mailbox Database Logs)
T: 50GB (Transport queue)

And yes, I use a separate VMDK for each drive. Silly to have partitions on virtual disks which makes disk expansions harder or unfeasible when volumes share disks. Plus it isolates the IO for each purpose. Exchange drives are formatted using 64k blocks.

With that said, this is all general Exchange-based advise an not necessarily Veeam-specific which I assume is what OP was asking for.

[dloseke]

Which logs from Exchange Server do you consider taking the Backup?

What this means is the mailbox database logs, similar to the transaction logs on a SQL server. When you are not using circular logging (I recommend not using circular logging unless you're in a pinch for space and need it temporarily to truncate the logs), Exchange generates log files for each database. These can be found in your mailbox database properties using the following EMS Powershell command that should list all databases from I believe all Exchange servers in your environment plus some other helpful information. If it only pulls from the server you're on, you can add a "Get-ExchangeServer | " to the front of this command -- I don't have any multiple-server environments to test against right now as almost all of my environments have been migrated to M365 because of the security exposure that Exchange presents in this day and age, not to mention less maintenance overhead. The mailbox database logs will reside in the LogFolderPath location assuming that circular logging is disabled. When Veeam and other Exchange-aware backup utilities complete a backup, it will signal Exchange to truncate the logs to what it has backed up and remove the log files that are no longer needed for a point in time recovery.

Code: Select all

Get-MailboxDatabase -Status | Select-Object Server,Name,EdbFilePath,LogFolderPath,CircularLoggingEnabled,DatabaseSize,AvailableNewMailboxSpace

[dloseke]

Also, most of the information you're going to find here is going to be older, but most of it still applicable and worth reviewing. As previously noted, check back with any other questions after reading.

https://www.veeam.com/blog/exchange-bac ... overy.html
https://helpcenter.veeam.com/docs/backu ... on_vm.html
https://bp.veeam.com/vbr/4_Operations/O ... hange.html

[karsten123]

better use official sizing calculator than guessing crazy numbers

[MPECSInc]

This sizing is...large...possibly. I mean, some of it makes sense, but for small environments with say, 50 users or something like that may require much small disks. For instance, I have one client running on a 50GB log drive. Assuming nightly (or more frequent) backups, and based on the amount of mail flowing through the server, this may be more than sufficient. I'm not sure what the 250GB disk is needed for on the install drive unless you're trying to account for the transport queue, but that can also be relocated. Exchange database drives are as needed based on the database sizes. My point is you're speaking in absolutes, and there's often not a one-size-fits-all sort of solution for this, so without knowing more about what the users requirements are, it's hard to make generalized statements here.

For me personally, but depending again on the size requirements of the users and expected mail flow, but this is where I start unless I know I'm going to need more....
C: 100GB (OS)
D: 100GB (Exchange Server Install)
M: 250GB (Mailbox Databases)
L: 50GB (Mailbox Database Logs)
T: 50GB (Transport queue)

And yes, I use a separate VMDK for each drive. Silly to have partitions on virtual disks which makes disk expansions harder or unfeasible when volumes share disks. Plus it isolates the IO for each purpose. Exchange drives are formatted using 64k blocks.

With that said, this is all general Exchange-based advise an not necessarily Veeam-specific which I assume is what OP was asking for.

If the OP prefers I can remove my posts. OP?

There are lots of reasons for starting where we do. The above is an example not gospel of what one can do. It is indeed advice.

Again, OP please indicate what you prefer and I will abide by that. Thank you.

[karsten123]

you forgot the restore partition