One-to-One Replicate Datastore on secondary storage

[cbraafhart]

Dear Forums,

I am currently trying to figure out what my options are seeing the method I was using is not exactly doing what I really want it to do.

I have 2 new ESXI hosts and i have 2 HPE iSCSI storages sitting on 10GB connections.

i was replicating storage enclosures to each other using the HPE direct replication method, but out of the box(confirmed with hpe) the replicated volume can not be used as in the way i thought i could have used it.

My goal is as followed:

I want to have a 1:1 copy from storage A to storage B
Storage A being my main production storage and Storage B being the if all fails storage.

I was thinking of using the COPY file job option, but after thinking it through, this method would be very inefficient due to constantly having to copy full the VM information, size and data every time i want to run the job. I thought about replicating the datastore complete(from Datastore A to Datastore B) but i really don't know if this is possible without having to actually mount or have the VM information displayed from Host A to Host during the replication process.

Does anyone have any suggestion what's the best method to do so and if other information is needed i will gladly provide it.

[Mildur]

Hi Christopher

I was thinking of using the COPY file job option

A file copy job would copy the same vmdks each day.

I want to have a 1:1 copy from storage A to storage B
Storage A being my main production storage and Storage B being the if all fails storage.

What sort of HPE storage have you in your environment?
If our product has support for your storage model, then you can do snapshot orchestration within Veeam.
Veeam will be aware of the VMs inside the protected snapshots, but not VmWare. If your first storage is failing, you have to recover the VMs with an instant vm recovery session. There should be no big performance penalty, because they will run directly from a clone on the second HPE storage snapshot. Depending you sized that storage similar to the first HPE storage.

Another option would be to use Veeam Replica or CDP Jobs instead of Snapshots.
If your first storage fails, just failover to the replicas on the second storage until you have solved the issue with the first storage.
When it is resolved, you can failback to the first storage. All changes on the replica will be replicated back.

If you want todo it without any reconfiguration, probably you have to build a metro cluster/stretched storage cluster between the two storages and esxi hosts. But I'm not an expert in that topic. Your HPE Sales/Engineers should be able to help you with that request.

Thanks
Fabian

[Andreas Neufert]

Veeam VM Replication with a planned Failover is the way to go, if VMware Storage VMotion can not be used.

[cbraafhart]

Dear Fabian and thank you for your replies,

We currently have 2x HPE MSA 2062 isci SAN.
Our storages are not in the supported model so therefor i can't use the snapshot orchestration I presume.
I will have to look into CDP Jobs(might be a different licensing schema?, currently have the enterprise License)

" If your first storage fails, just failover to the replicas on the second storage until you have solved the issue with the first storage.
When it is resolved, you can failback to the first storage. All changes on the replica will be replicated back."

And this is exactly what i was trying to achieve with the SAN storages and the built in replication capability, but it was creating a replicated vmfs volume without being able to actually mount or be seen to my ESXI host, so back to the drawing board I said. Therefor i am currently looking for alternatives that could meet the goal i am trying to achieve.

Nothing fancy just the virtual machine data and information should be replicated(hourly preferably) that if all fails i can just boot up and us immediately just as you said.

"Veeam VM Replication with a planned Failover is the way to go, if VMware Storage VMotion can not be used" Will also check this, was checking this yesterday but it got at a point i needed to go to sleep.

[Andreas Neufert]

The VM Replication Jona with an hourly schedule and prepared Failover Plans (create them in addition to the VM Replication Jobs) is I think what you need based on your existing license model. Enterprise Plus could give you near continuous replication options if you have as well a vcenter.

As well I would explore if the storage supports an active active volume spanned over two storages (mirroring). This way both sides serve the same workload.