Exchange DAG Backup and Recover: The right way!

[ajped]

Hello,

I just starting to backup my Exchange Infrastructure with V&R 10. We have two nodes (virtual machines) with active databases in both nodes. I would like not wasting space so instead backup both nodes everyday, I have created two different task and I backup a node one day, and the other one the following day. Besides, I make a full backup in the even months from one node and a full backup from the other node in the odd months.
Now, I have read https://helpcenter.veeam.com/docs/backu ... ml?ver=100 and It says: Veeam Backup & Replication supports any configuration of DAGs. In particular, having all the databases active on a single node or with the active databases on each node. Transaction logs will be truncated on all DAG members.
Because of "Transaction logs will be truncated on all DAG members", I don't know If I am backing DAG in the right way. If we backup the nodes in different tasks and different days, Could we have any problem when we will make a restore, or this doesn't affect?.

Instead of our aproach, It will be better to backup both nodes in the same task, or maybe backup always only the same node?

Thank you.

[Andreas Neufert]

For the log truncation part: Veeam itself do not decide if the logs will be truncated and on what systems with what replication. We tell the VSS framework that we are ready with backup and the Exchange writer will decide what to do.

In case of backup saving, if you prepare well you can backup both DAG nodes in the same job and use per Job chains for the deduplication. That way you can backup both servers full and let the deduplication reduce the data.

In general if you only backup active or passive databases you have always the issue with the human factor which can select the wrong stuff. As well at restore it can get complicated. What to restore and how to establish replication again. We had lately some customers that struggeled with the recovery after a broken Exchange Update. If they would have backed up the full server, it would have been a very simple approach (just restore the broken server to latest state and Exchange will be put in recover mode and data get´s replicated). But they ended up with a lot of preparation, Forum questions, support calls until they had the clear picture for the restore. I know it is space and it cost, but in case of disaster recovery, to have everything backed up is so simple to do, specifically if you are not a Exchange expert.

Check out this one for the preapration of the cluster to make sure the cluster will not failover at VMware Snapshot processing, storage vmotion,... https://andyandthevms.com/exchange-dag- ... plication/

[ajped]

Hello Andreas, thank you for your answer.

I have read https://andyandthevms.com/exchange-dag- ... plication/ and It says:

For DAG clusters there are 2 ways in case of Job design:
– Repository: Per Job chain with all DAG members in same job.
– One DAG member per Job
The idea of the first one is to use Veeam Inline Deduplication to reduce data amount significant (as the DAG members have a lot of replicated data). But in case you run into Snapshot commit issues with Cluster failovers, this can be negative as it is likely that shortly after a cluster failover the next snapshot commit happens on the other member. The result would be double cluster failover within short time. In theory this should not be a problem, but I saw customers struggling with it because the index service consume then 100% of the CPU for a long time and User connections timed out. If you want to be 100% on save side or use per VM chain Repository setting anyway, create one Job per DAG member (or site) and schedule the Jobs at different times within the night.

In my case, I have created one Job per DAG member but I schedule the Jobs at different days and not in the same night (one day a node and the other day the second node). I backup one node everyday and I make full backup from the same node every two months (one month a make a full backup from a node and the following month the other node). So, I suppose I am backing active and passive databases but I don't know It's the right way.

I'm worried about the space because I backup to disk repository (and here I have no problem with the space because I have x recovery points) and to tape (for archival purposes and here I would like to save space).

So, in your opinion It's better to create only a task with the two nodes (and enable inline data deduplication) or one task per node in different days?

Note: I have increased the DAG heartbeat time to avoid cluster fail over but I have not tested to back both nodes at the same time.

Thank you again!

[Andreas Neufert]

If you use Veeam, it do not matter if you backup active or passive databases, as our Veeam Explorer for Exchange can work with both.
Usually customers try to backup the passive databases as the load is reduced there (only writes for the DB replication but no reads from the clients).

If you have stable backup processing without cluster failover at snapshot commit time, there is no issue with backing up the 2 systems in one job.
The scenario to avoid is that you have multiple cluster failovers during backup as the index service would go crazy in case of CPU consumption.
I did not tested the Failover/Index Service correlation with newer than 2016 Exchange versions, but I assume it is the same there.

[ajped]

Hello Andreas and thank you again.

I have tried twice to backup two nodes in the same backup task and I got the warning message: VSS: Backup job failed.
Cannot notify writers about the 'BACKUP FINISH' event. A VSS critical writer has failed.... Writer's state: [VSS_WS_FAILED_AT_BACKUP_COMPLETE]. Error code: [0x800423f3]. This message is always for the node whose backup ends after. Although I haven't seen any cluster failover and the database keep active in the same nodes that before the task started.

So, I'm still not sure what is the best aproach. What's more, I don't know why, but with one task and the two node backup I haven't seen reduce data amount significant (It's almost double that backup only one node).

Anyway, I have just opened a support case with Veeam.

Thank you!

[Andreas Neufert]

So you have a IP less DAG?