Backing up Always On MSSQL with Veeam

[agentpineapple]

Hi Vitaly,

So reading the thread, Veam supports backing up the active node in a SQL always on cluster?

But does Veam also support replication of the node?

Thanks

[Vitaliy S.]

Hi Dan, you can either backup and replicate these VMs with application-aware image processing option disabled or if you have only active database present on the VM, you can enable this option in the jobs.

[veremin]

And to backup VMs hosting passive replicas some of our customers leverages the following workaround outside production hours:

1) Create a script that turns passive node into active one, performing manual failover.
2) Execute a backup job.
3) Once the snapshot is taken, reverts everything back.

Thanks.

[Gostev]

Just wanted to add that we are working on adding native support for application-aware backup of AlwaysOn SQL servers without the need to use scripts. Thanks everyone for providing your feedback, this helped to raise the priority of this feature.

[Alek14]

Hi,

I'm new to Veeam product.
I'm trying to backup AlwaysOn SQL and followed advise above to make the backup of Passive Node. The backup runs fine but I'm seeing that Data and Log Drives which are shared VHDX are being excluded. As these are the drives I really need backing up I'm wondering if I'm missing something here?

Thanks,

Alek

[foggy]

Alek, shared VHDX disks are currently not supported for backup by Microsoft, and hence automatically excluded from backup.

[Alek14]

Ok...so how do I effectively backup an SQL DB (AlwaysOn) then? thanks

[Gostev]

I am afraid there is nothing you can do until Microsoft adds support for backing up shared VHDX. A lot of people are frustrated with this limitation, since obviously lack of backup turns this whole shared VHDX feature into unusable gimmick, so I am sure this feature is a priority for Microsoft.

[mdornfeld]

Add me to the list of people who are very interested in some kind of Best Practice whitepaper or KB article on how to correctly backup SQL 2012 Always-On clusters on the VMware platform using Veeam. I did open a case #00582605 to request the same directly from Veeam support as well. Thanks.

[veremin]

As mentioned above, we're working on adding native support for Always On SQL. Chances are, sooner rather than later there will be no need in KB article, as everything will be done automatically by Application Aware Image Processing. Thanks.

[lightsout]

I actually have my vCenter DB on the Always On cluster too. Veeam correctly identifies that it is the VMware DB, however, it uses the name of the listener rather than the name of the VM. So I had to do this:

http://www.veeam.com/kb1051

To allow me to backup the active node again.

[enettekoven]

I am having the exact same issue. In our situation the AlwaysOn SQL is used to host the SharePoint Databases. As soon as there is an update, will that be posted here or is there some other way to be kept in this/the loop?

[Vitaliy S.]

Yes, you can either subscribe to these forums or check our What's New document that is always shipped with our updates/releases. Thanks!

[enettekoven]

This topic only describes the backup of the alwaysOn, how about the restore especially when hosting SharePoint DB's and trying to use the Veeam Explorer for SharePoint?

[Vitaliy S.]

Hi Erik,

Can you please clarify your question? Do you see issues with using Explorer for SharePoint and Always-On SQL Server databases?

Thanks!

[mikegmi]

Likewise, please let me know when you have released a whitepaper with guidelines, or have a way to backup SQL AlwaysOn databases natively. We too are experiencing issues.

[Dave23]

I know that Veeam doesn't support replicating SQL 2012 Availability groups, but has anyone found a way to make this work? I'm able to replicate both nodes in my cluster from my production to DR site (that are on two different networks), but when I failover and bring both replicas online, the Availability group is no synchronized and neither are the databases because the databases are not in a healthy state. I also have to go into the Failover Cluster Manager and modify the AG listener with a new IP address since the cluster is now on a different network. If anyone has any input on how to make this work that would be great.

Thanks,

Dave

[lightsout]

I applied SQL 2012 SP2 to my machines, and now it seems, I can backup the secondary node. Has anyone else seen this?

[veremin]

Can you elaborate on what replica the given VM hosts? Only active, only passive, or both active and passive? Thanks.

[lightsout]

Both active & passive nodes are VMs. I only backup one, as it only works if it is the active node.

The ultimate problem was a SQL VSS writer would error on a passive node, causing the failure. I would say that seems to be fixed with SQL 2012 SP2 on one of my clusters. I'm upgrading the other to see if it helps that one further.

[veremin]

I think we haven't had issues with backing up VMs that host only active node. Or you're saying that you experienced different behavior that have been changed by applying SP2? Thanks.

[lightsout]

Yes, that is what I'm saying. It seems to be the case I can backup passive hosts running SP2.

Not all of my SQL Servers are upgraded yet, so I'm in that process now, but so far it seems to have resolved it.

[tsightler]

I suspect the following change from the SQL2012 SP2 change notes is the reason:

Functionality:
SQL Writer support COPY_ONLY backup: Data Protection Manager (DPM) full backup on the server with AlwaysOn secondary will be converted to COPY_ONLY backup.

Obviously Veeam <> DPM, but we're using the VSS SQL Writer as well and it sounds like this would convert any call to perform a VSS freeze for the database on a passive node to a COPY_ONLY call (i.e. no log truncation), which should work fine for us as well.

[vmstrata]

Hey guys. Any updates on this issue?

[lightsout]

All of my AAG SQL 2012 servers are updated to SP2, and the issues seems to be resolved for me.

[foggy]

We are adding full support for Availability Groups backup in Veeam B&R v8, stay tuned.

[sjlarsen]

We are adding full support for Availability Groups backup in Veeam B&R v8, stay tuned.

Any updates on when?
Just updated to v8, and still have this problem:

Code: Select all

21-11-2014 07:05:41 :: Processing SQL04 Error: Unfreeze error: [Backup job failed.
Cannot create a shadow copy of the volumes containing writer's data.
A VSS critical writer has failed. Writer name: [SqlServerWriter]. Class ID: [{a65faa63-5ea8-4ebc-9dbd-a0c4db26912a}]. Instance ID: [{e7de7332-63c5-401e-b235-19956f55865b}]. Writer's state: [VSS_WS_FAILED_AT_PREPARE_SNAPSHOT]. Error code: [0x800423f4].]

[Gostev]

v8 has added full support for SQL Always On (both at backup and restore). Please, open a support case.

[igarcia1]

The new user guide doesn't really mention, but seems to infer that all the AG nodes need to be in the same job. We have four nodes in the cluster that spans two sites, two nodes at each sites, for which we use separate repositories and vCenter deployments. I started setting up two different jobs for each pair of nodes to backup to each site, but started wondering about the implementation.

This should be okay to do as long as the backup preference is to use the secondary node, correct? The end result would be duplicate backup data (for the databases that span both sites) and truncated logs across the board, no?

[foggy]

AG nodes do not necessarily need to be backed up in a single job. What is much more important, is that transaction logs backup should be performed only in one job - if some of the nodes are backed up by a different job, this job should have the 'Do not truncate' setting selected for these VMs.

I agree that user guide should cover that better, will forward your request to the documentation team. Thanks!