Exchange DAG - DR Scenario

Question

question

HildebrandtNiclas-8679 asked Apr 26, '22 HildebrandtNiclas-8679 edited Apr 29, '22

Exchange DAG - DR Scenario

Hi all,

I am currently testing our desaster scenario and try to make a documentation.

Scenario:
One DAG containing 3 databases (2x Exchange 2016 on site, 1x Exchange 2016 in Datacenter)

Exch01 (on site) holds DB01 and DB02, Exch03 (on site) holds DB03.
Exch02 located in the datacenter holds a DB copy of all DBs

In case of desaster (fire, war, bombing the city, etc), the on site hardware is lost and we want to keep services up using the hardware in the datacenter.

As 2/3 servers are shut down abruptly, the DAG will not shift the active mount to the datacenter and additional the quorum of the DAG is lost.

Now the race begins:
1. To regain quorum I would need to remove one server from the DAG (so the FS-witness would take over again), but I cannot remove any server as it "holds" the active copy of at least one DB
2. To get rid of that active DB, I would need to remove the DB Copy, what is not possible, as it is marked as active
3. To deactivate the DB I would need to move it to another DAG member, what is not possible, as DAG quorum is not available and cluster is down

At the current point I cannot find the correct way how to start.

Having two exchange in both sites would result in a FW witness that either sits on site or in the datacenter and whenever the WAN is gone, our DAG would unmount the DBs -> no solution.

Looking forward to any answer that may guide me to the correct startpoint.

Brgds, Niclas

office-exchange-server-connectivity office-exchange-server-ha

· 1

5 |1600 characters needed characters left characters exceeded

Attachments: Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

KyleXu-MSFT · Apr 27 at 07:48 AM

@HildebrandtNiclas-8679

From this blog, Dynamic Quorum could prevent this phenomenon occurs, I think it may be a good choice for you.
Please Note: Since the web site is not hosted by Microsoft, the link may change without notice. Microsoft does not guarantee the accuracy of this information.

You can also follow the article that was provided by AndyDavid to do a datacenter switchover.

0 ·

Answer 1 · 2022-04-26T15:09:26.867Z

AndyDavid answered Apr 26, '22

Enable and test datacenter switchovers. This will guide you:

https://docs.microsoft.com/en-us/exchange/high-availability/manage-ha/datacenter-switchovers?view=exchserver-2019

You have to enable DAC mode for this to work smoothly:
https://docs.microsoft.com/en-us/exchange/high-availability/database-availability-groups/dac-mode?view=exchserver-2019

5 |1600 characters needed characters left characters exceeded

Attachments: Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Answer 2 · 2022-04-28T14:38:03.24Z

HildebrandtNiclas-8679 answered Apr 28, '22 AndyDavid commented Apr 28, '22

Hi,

i already read about it and alsoenabled it. During the test of stopping the DAG it also tried to connect to a failed member, thus the procedure stopped.
I assume I need to set verything up with different AD sites in order to have it working properly?

I busted my testing environment and need to reinstall everything, but will test it again afterwards.

· 1

5 |1600 characters needed characters left characters exceeded

Attachments: Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

AndyDavid · Apr 28 at 02:49 PM

ITs designed for 2 different data centers. Typically that would be 2 diff sites. yes.

0 ·

Answer 3 · 2022-04-29T18:47:17.867Z

HildebrandtNiclas-8679 answered Apr 29, '22 HildebrandtNiclas-8679 edited Apr 29, '22

Hi again,

I resintalled the test lab.

Once two Exch are powered off (that had the acive DB copy), the remaining one does not get active.

As DAC is enabled, I then try to mark the stop the DAG fpr the failed servers, but this also does not work:

[PS] C:\Windows\system32>Stop-DatabaseAvailabilityGroup -Identity DAG -MailboxServer EXCH01

Confirm
Are you sure you want to perform this action?
Stopping mailbox server "EXCH01" in database availability group "DAG".
[Y] Yes [A] Yes to All [N] No [L] No to All [?] Help (default is "Y"): a
An error occurred while attempting a cluster operation. Error: Cluster API failed: "GetClusterKey() failed with 0x46.
Error: The remote server has been paused or is in the process of being started"
+ CategoryInfo : NotSpecified: (:) [Stop-DatabaseAvailabilityGroup], ClusterApiException
+ FullyQualifiedErrorId : [Server=EXCH02,RequestId=5d1f0b3c-489d-4055-bb3b-e60ce3a424af,TimeStamp=29/04/2022 18:34 :49] [FailureCategory=Cmdlet-ClusterApiException] 36474480,Microsoft.Exchange.Management.SystemConfigurationTasks.
StopDatabaseAvailabilityGroup
+ PSComputerName : exch02.test.domain

All servers are currently in the same AD site.

€dit: I just found my problem in not usgin the -configurationonly parameter.

With that in place I was able to fianlly shut the member down and restore the DAG with alternate FSW.

DBs were mounted automatically!

5 |1600 characters needed characters left characters exceeded

Attachments: Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

question