Slightly complicated replication

[nixon]

Hi all,

We have the following scenario:
Prod Cluster A\Server 1 -> replicates across a WAN link -> DR Cluster B\Server1

The issue that we are having is that the size of the server and the width of the pipe means that we are running into issues with the [separate product] backup windows.

As an idea, we are looking at doing the following:
Prod Cluster A\Server 1 -> Replication job within the cluster over the LAN -> Then replicate that replica to DR Cluster B\Server1

Does anyone else have this sort of setup in place? The main question is regarding the transition. Will we be able to use the existing DR Cluster B\Server1 virtual machine as a target for the LAN replicated server (as opposed to the original server) using replica seeding?

Just to rephrase if the above is not clear:
Currently: Server A replicates via WAN to server Server Z
Want to: Replicate Server A via LAN to Server B, then replicate Server B to Server Z
Do we need to start again with the offline seeding, or will this work?

PS: The policy of not dealing with support issues in the forum, and then having a log in only KB with a really poor search engine is a stupid stupid idea. Assign some Veeam people to look after forum issues - if they are complicated -send to support. Not being able to look at other peoples issues wastes your time and mine. Make the KB public, and put some more articles in it...

Thanks

[Jfmoots]

Just to be clear...

You want to replicate server A over the LAN to Server B, then Replicate Server B to the EXISTING Server Z. I understand you're having trouble with the backup window, but I'm trying to wrap my head around how this helps. How does doing this help you?

In the end, you want to use Change Block Tracking to send changes from Server A to Server B? Then from Server B to Server Z? I'm assuming those will be on some sort of different schedule and that's what's helping your backup window? Failover/failback becomes interesting to. What's your plan there?

I'll try what you're doing in my lab early this week and let you know if I can map an existing replica to a replica of the original VM. I've never tried it and don't know what the result will be, but I'll let you know.

[Jfmoots]

I didn't think about something...

"2/26/2012 9:59:11 AM :: Changed block tracking cannot be enabled: one or more snapshots present."

When I tried to run a job from the LAN Replica to the WAN Replica, I got this message. We store our restore points as Snapshots. You can't leverage CBT on a VM with snapshots.

I was able to map the LAN Replica to the already existing WAN Replica without an issue.The job ran fine, but it didn't leverage CBT.

[Gostev]

You can't leverage CBT on a VM with snapshots.

You can, but CBT must be enabled on the VM when it has no snapshots. Meaning that you should do this (either manually, or with a help of Veeam job) after the 1st replication job run, when replica VM does not yet have any snapshots.

[zion]

[merged]

Good Day,

We are testing a Veeam implementation. The setup is as follows:

Server1 on ESX1 at Site A replicating to ESX2 at Site B controlled by VeeamServer1

Replica of Server1 on ESX2 at Site B replicating to ESX3 at Site C controlled by VeeamServer2

We were testing this a while back before Version 6 came out and the problem we had is like this.

A single replication job from Site A to Site B works fine, we can test failover and see that the failed over replica is up to date as of the last scheduled replication.

If we then replicate that replica to Site C it will also be up to date, however, any changes to the VM, while they are carried over to Site B do not carry over to Site C.

As I said we were testing this with Veeam5. Is this an implementation we can have with Veeam6? Is this even possible? What I was thinking was that the Veeam looks for changes made by an ESX server, however after it's replicated, the subsequent changes are made/saved by Veeam and as such when another Veeam server looks for "VMware" changes, it doesn't see them.

I don't know if that is right but I would appreciate any and all input.

Regards,
Zion

[zion]

I see I have been merged into this topic; however I don't think my issue is the same. We had no problems with the scheduling. The jobs all "complete successfully" as far as Veeam reports. It's just that at the final destination of the replica (the replica of a replica) the changes do not carry over. Maybe I don't understand properly and the merger into this thread is relevant, if so could someone please explain?

Regards,
Zion

[Vitaliy S.]

Hi Zion,

Yes, I agree that this topic does not explain your issue.

Here is the relevant one: Backup the replicated VMs (it applies to both backup and replication jobs)

Thanks!