Veeam VAOv3/Netapp MirrorVault

[stidedo]

Hi,

We want to implement this setup and I was wondering if vao v3 is supporting this at the moment:

• Configure a VBR job on our VBR server that's triggering a Netapp MirrorVault relation between netappHQ and netappDR. (we want to use MirrorVault relation because we want 1 data copy from source to destination Netapp, we don't want to setup 2 separate jobs snapmirror/snapvault)

• Configure a VAO failover plan on our VAO server, that:
  • break snapmirror(MirrorVault) relation between netappHQ and netappDR
  • register vm's on netappDR in vCenter on the DR ESX host
  • re-ip vm's
  • power on vm's and execute some custom powershell scripts to start applications correctly

• Possibility to snapmirror data back from netappDR to netappHQ afterwards

Can we implement this scenario fully automated from within VAO v3? Or are there still parts we have to execute manually before we can run our VAO plan? Is MirrorVault supported or should be configure a separate snapmirror relation between netappHQ/netappDR?

Thanks.

[Alec King]

Hello,
Yes, your scenario is fully supported by VAO orchestration
VAO v3 supports both Mirror and MirrorAndVault NetApp relationships.
VAO can automatically perform all the tasks you list above (break NetApp relationship and mount volumes on DR, register VMs on DR, re-IP VMs, power on and execute additional steps/checks/scripts)
VAO can also trigger the reverse-resync in NetApp to prepare for failing back to production. (you may need to modify your VBR job to restart the replication from DR to HQ, this depends on your specific configuration)
And of course, the testing of this DR plan can also be automated in a VAO DataLab.

Hope that helps - any further questions please let us know!

[Cyril]

Hello,

In the same way, VAO support the relationship for SVM DR ?

Thanks.

[Alec King]

Hi,
VAO does not directly orchestrate SVM DR - in current version we manage volume failover directly.
Although SVM DR is not currently supported, VAO will autodetect all the vSphere datastores which are backed by NetApp storage when the SVM is added. Therefore the VAO Recovery Location will automatically be configured with all required volumes. This will achieve the same result as full SVM failover, but by orchestrating failover per-volume instead of per-SVM.

[Cyril]

Ok. For a particular case, we have SVM in NFS with exports for Datastore and exports for anothers applications in this VMs. The SVM DR bring the simplicity that the IP address stay the same for the NFS in the VMs. Pity...

[Alec King]

Ah, I see. There may still be workaround, for example some scripting within VMs to change IP for NFS application drives? VAO can trigger such scripts to run inside VM guestOS, with different parameters depending on failover, failback, etc.

[Cyril]

Thank you, i will watch this.

[dries.vergote]

Hi

In addition on the original post, I'm also involved in the configuration.
The make leverage of the netapp in VAO, do you need to add the SVM to VBR as well or is adding the cluster of netapp also possible?
In VAO you only can add the SVM.

In the report we see that the datastore has no backups and also in the plan we are selecting a storage system SVM but there are no Datastore in the "datastores backed up with this storage system" menu.
We had added the netapp as cluster to the VBR and created a vm job with snapvault config and that is working.

Thanks in advance

[Alec King]

Hi,
You must add all NetApp SVMs to VAO (both the source and target SVMs) which are protecting volumes where vSphere datastores are/will be mounted.
You also must add the vCenters where these datastores are used.

It's optional to add the Cluster to your remote VBR server. VAO doesn't require this. Of course there are benefits to driving snapshot jobs from VBR, e.g. app-aware Guest processing etc.

It's also optional to deploy VAO agent to the remote VBR, as VAO can orchestrate failover with only the above SVM + vCenter connections. However it is a good idea to deploy the agent, so VAO can gather the configuration, restore points available etc.

Just to confirm - have you created a VAO plan of 'Storage' type? Have you also created a Storage Recovery Location in VAO administration? You should see Replicated Datastores under the SVM(s) for each configured recovery location.

Hope that helps!

[dries.vergote]

Hi Alec

Thanks for your answer.
In VAO:
We have added the vcenter, default vbr server installed with VAO, all our netapp SVM's.
If we select a storage recovery location to the destination netapp svm, we don't see any datastore there. Also not on the primary datastore.

We have a backup job running in the VBR on the same netapp SVM's that are in VAO.
The backupjob is netapp snapshot only with snapshot-vault to secundary netapp.
The datastore is setup in Mirror-Vault relationship config. So there is as well a mirror relationship.

Why aren't we seeing any datastores in "datastores backed up with this storage system" menu.
Where does he gather that information from? As there are veeam snapshot's and netapp mirror snapshots on the drive.

[Alec King]

Hi Dries,
VAO will build the 'map' of NetApp volumes -> vSphere datastores using its vCenter and SVM connections. If you don't see any datastores in your VAO recovery location, and everything is connected correctly, there could be several reasons (e.g. DNS issues, permissions of the vsadmin account etc)
I suggest you open a support case on our customer portal - the support team will be able to check your configuration in detail.

[stidedo]

Hi Alec,

We're facing another issue ( case 04456027):

• On Netapp we have different data lifs we are using for NFS datastores. We are not using DNS, it's ip based.

• On our ESXi hosts we are using a fqdn in the hosts file to mount the NFS datastores

As I understand VAO will use the vCenter MOB to search "remotehost". In our case remotehost = storage-dc1-01 (defined in the hosts file). On Netapp this name is not known. VAO sees an IP. This IP doesn't match with the remotehost name of vcenter. So VAO decides there's no snapmirror relationship for that datastore. We already tried to add "storage-dc1-01" to the hosts file of the SVM. But stull VAO can't match this.

How does VAO verify "remotehost" on Netapp? Is there a way to fix this? Behind storage-dc1-01 we have mounted a lot of datastores. So it's not really an option to power off all vm's/remove from vcenter inventory/remount datastore on ip/power on vm's for all customer vm's/datastores?

Thanks

[Alec King]

Hello,
According to our research you can configure DNS on the NetApp SVM as follows -

Code: Select all

network interface modify -vserver %vserver_name% -lif %lif_name% -dns-zone %domain_name%

VAO will then insert appropriate DNS entry into it's own database, this will allow it to detect storage mappings correctly.

[chjones]

Hi Alec

Thanks for your answer.
In VAO:
We have added the vcenter, default vbr server installed with VAO, all our netapp SVM's.
If we select a storage recovery location to the destination netapp svm, we don't see any datastore there. Also not on the primary datastore.

We have a backup job running in the VBR on the same netapp SVM's that are in VAO.
The backupjob is netapp snapshot only with snapshot-vault to secundary netapp.
The datastore is setup in Mirror-Vault relationship config. So there is as well a mirror relationship.

Why aren't we seeing any datastores in "datastores backed up with this storage system" menu.
Where does he gather that information from? As there are veeam snapshot's and netapp mirror snapshots on the drive.

Did you ever find a solution for this? I’m trying to build this in a lab using Veeam v11 and DR Orchestrator 5 and get the same issue of the datastore showing it has no protection.

[Alec King]

Hi Chris,
This is a core supported scenario for Veeam Orchestrator, it should definitely work. I am looking at this exact configuration in my own lab right now
The central requirement is to connect ALL the storage nodes and the vCenter to VDRO; and deploy VDRO agent to the Backup server too (if using a remote one)

It can take a little time for the information across all those systems to be collected and synchronised in VDRO. You can check the status of the storage nodes in the Storage Systems view in Administration section.
Have you been able to create a storage Plan? If so the Readiness Check (or Quick Check) will give you information regarding the replication status of the datastore.

If you are still having issues, please open a support case - we can then dive into your logfiles. Thanks!

[chjones]

Hi Alec, thanks for the reply. I managed to fix this myself. The SnapMirror Policy I was using was a "Vault" type. Switching it to a Mirror or Mirror-Vault policy and ensuring SnapMirror was one of the secondary destinations in the Veeam B&R Job did the trick.

Veeam DRO is immediately aware of the relationship as long as Mirror is one of the destinations, not just a Vault.