Using Veeam replication to Migrate Datacenter

[priusit]

I am considering on using Veeam Backup & replication to migrate a datacenter to a new location. Has anybody done this before and what are the steps that I have to take?

What issues and pitfalls are there?

I hope that somebody can help me on this

[Vitaliy S.]

Hi Cor,

You can either use replication or Quick Migration jobs to do that. Please review our Evaluator's Guide for more info.

As to possible tips and tricks for VM migration process, then it would help if we could know a bit more on your environment

Thanks!

[dellock6]

Hi Cor,
I did it. Last was for a customer moving all his racks in a datacenter, 300km away, and we used replication jobs. Issues and pitfalls are not related to Veeam but more in general to network setup, here are some points you need to check/consider:
- networking MUST be completely correct, routing and dns above all
- if you are going to connect via L3, check with some replica jobs re-ip of VMs is working
- if you can, use replica seeding to save much time and bandwidth, our migration was 4Tb of VMDK disks...
- if you are going to keep both sites in active state, prepare a list of VMs based on priority to be moved, dependencies between VMs and such

But most important of all, be sure you can do failback if something is not working.

[scooke]

Hi guys, I figure this is the place I'll get the best answer for this question.

I am about to receive and entirely new infrastructure, New hosts, new SAN, new vCentre Server (VMware Essentials Plus 5.1)
My old environment is VMware Enterprise 4.0

What is the best way to move the VMs from the old infrastructure to the new one with the least downtime?

I would say a replication job in Veeam B+R but as far as I've tried, the VMs all go into a Veeam folder on the SAN, IE "Datastore1\Veeam\Server1", and I'd rather have it as "Datastore1\Server1" on the datastore.

I know I can just power off the server and do a fastSCP but that will be significant downtime.
I also thought of bringing the SAN and hosts into the old environment and doing a Storage and host migrate, but the new SAN is SAS and the old one is iSCSI, and also I don't have licenses for more hosts in the old environment.

Any ideas will be greatly appreciated

Thanks
Steve

[kjc3303]

Hi Steve you can use Quick Migration, See below for description

You should also be able to achieve Datastore1\Server1 I have in my environment and don't remember having to do anything unusual?

Cheers

Kev

Veeam Quick Migration enables you to promptly migrate one or more VMs between ESX(i) hosts and datastores. Veeam Backup & Replication allows migration of VMs in any state with minimum disruption to business operations and end user access to services. You can use Quick Migration as a self‑contained capability, solely for VM migration, or combine it with Instant VM Recovery.

Veeam Backup & Replication analyzes your virtual environment, its configuration, the state of VMs and selects the most appropriate relocation method. Whenever possible, Veeam Backup & Replicationcoordinates its operations with vCenter Server and uses native VMware vCenter migration mechanisms: vMotion and Storage vMotion. When VMware vCenter migration methods cannot be used (for example, if your VMware vSphere license does not provide support for vMotion and Storage vMotion, or you need to migrate VMs from one standalone ESX(i) host to another), Veeam Backup & Replication uses its proprietary SmartSwitch technology to relocate VMs.

Veeam Quick Migration provides means for fast background migration of VMs ensuring continuous uptime of your virtual environment. Quick Migration supports hot VM migration (with SmartSwitch) and cold VM migration (with cold switch).

Migration of a VM is performed in several stages:

1.Veeam Backup & Replication copies VM configuration (.vmx) to the target host and registers the VM.

2.Veeam Backup & Replication triggers a VM snapshot and copies VM disk content to the new destination.

3.VM state and changes made after snapshot creation are moved to a new location. Veeam Backup & Replication uses different approaches to move the VM state between hosts with compatible and non‑compatible CPUs.

•If you move a VM between two hosts with compatible CPUs, Veeam Backup & Replication uses SmartSwitch (that is, it suspends a VM to move its state file and changes made after snapshot creation). The VM is then resumed on the new host. This ensures minimum downtime, and completely eliminates any data loss during migration.

•If you move a VM between two hosts with non‑compatible CPUs, Veeam Backup & Replication stops the VM to move changes made after snapshot creation, and then starts the VM on the new host.

Integration with Instant VM Recovery

When you restore a VM using Instant VM Recovery, Veeam Backup & Replication starts the VM directly from a compressed and deduplicated backup file. To finalize recovery of a VM, you still need to move it to a new location. Moving the VM with VMware Storage vMotion or hot replication may require a lot of time and resources, or it may cause loss of valuable data.

Veeam Quick Migration was designed to complement Instant VM Recovery. Instead of pulling data from vPower NFS datastore, Quick Migration registers the VM on the target host, restores the VM contents from the backup file located in the backup repository and synchronizes the VM restored from backup with the running VM.

[foggy]

Hi guys, I figure this is the place I'll get the best answer for this question.

Indeed!

I would say a replication job in Veeam B+R but as far as I've tried, the VMs all go into a Veeam folder on the SAN, IE "Datastore1\Veeam\Server1", and I'd rather have it as "Datastore1\Server1" on the datastore.

Please review the answers given above, should help in your scenario. As for the target folder for replicated VMs, then you can configure it in the replication job settings and specify the required one.

Thanks!

[scooke]

OK awesome, I didn't even know Quick Migration existed!

Thanks for the info guys, I will give it a go once I get the new kit and configure it!

[foggy]

Should any issues arise during setting up the migration process, feel free to ask here.

[jmerfeld]

I am trying to figure out a way to accomplish project I am going to work on.

We are going to have a remote site that is going to have terminal server for about 25 remote users to work on. We also have a terminal server at another location (hq site)where veeam sits. What my plan is.. Is to be able to patch the terminal server at the hq site and then send the server up to the remote site and use that after renaming re-ip etc and the remote site users can use that server.

My issue is how to get the server to the remote site the quickest. And easiest. I am unsure if a replication would be the best idea.. Or what.

There is a 20mb pipe between the two sites if that matters

Help?

[Vitaliy S.]

Hi James,

Yes, replication job is the way to go in this case. Install two proxy servers on both ends to send compressed VM data across the WAN link and run the replication job.

Thanks!

[jmerfeld]

Is it safe to delete the active snap on the other end when I power it up? What is going to be my issues with deleting the snap? Ot consolidating it.

[Vitaliy S.]

Yes, it is safe. When you plan to failover make sure you choose permanent failover option in the backup console to remove replica VM records from the console.

[dellock6]

And to further assure James, I did a similar task for a customer, we used Veeam replicas to move an entire vSphere cluster in another datacenter 400 kms aways, we had 150 server VMs in there and got really no problem at all. We used permanent failover for each VM.

Luca.

[jmerfeld]

I guess what we are after is being able to have a copy of a terminal server at site a, make changes to that machine, power off, replicate to site b, consolidate the snap and then power it back on and let users work from it at that site b (Thin Clients).

Does that make sense?

[dellock6]

Absolutely.
Is the same activites we did. Veeam replicas let you raiterate the same replica job again and again. You can start with a first replica (maybe also using seeding options), then replicate again to keep the failover VM in sync with the primary, and at the end do a final replica using failover command.

Luca.

[jmerfeld]

If I wanted to send the replica to site b.. then a month later after people work on it.. replicate it back to site a for updates.. make my changes, and then replicate it back to site b for people to work on again, I would need to license site b with veeam to pull this off correct?

or is there another way.

[foggy]

You could temporarily revoke the license from the source host and assign it to the host in site B that would serve as source for the replication job, then revoke and assign it back.

[sdelacruz]

Hello everyone,

I am in a middle of a debate. I have 3 hosts. 1 host is running Vsphere 4.1 esx. The other 2 hosts are running Vsphere 5.0 esxi
I am running Veeam V&R 6.5.
On my 4.1 host, I am currently running out of space on my datastore. I am in need to move a DC living on ESx 4.1 to one of my 5.0 host.
I want to move the DC using veeam but do not know which is the procedure to do it. Is there a best practice to do this? Can this be accomplished with veeam?

Any help is greatly appreciated.

SAM

[Vitaliy S.]

Hello Sam,

Yes, it can be accomplished with either replication or Quick Migration jobs. Moving DC should be the same as moving any other VM, just make sure you use AAIP (application aware image processing) during migration to the another host.

Thank you!

[veremin]

It might be still better to follow with replication scenario, since VM you’re going to migrate is DC and quick migration doesn’t utilize AAIP (Required for proper quiescing of such VM). Thanks.

[sjmp]

I want to move a VM in production to a WAN DR site. And was going to use Veeam to replicate the VM.

If I setup a replication job to move this. Once it is moved over can I remove it from Veeam and the vm from production and not effect the replicated VM in DR?

Basically I will not need to replicate this once the initial job is complete. Are there any instructions, recommendations or problems with this?

Thanks,

SJ

[foggy]

Scott, please review the thread above for similar discussions. Basically, you can safely remove the original VM after performing permanent failover and making sure the replica VM operates as required.

[royp]

We have a licenced copy of Veeam BnR vers 7.
We don’t have the Wan Accelerator Licenced, but would look into it if it helped.
We have Esxi 5.1
We have vSphere Storage Motion and Vsphere vMotion along with vSphere HA

Scenario is:
1 current site with a 2 meg Leased line which usually gives an average upload of 3.5 Mbps
Shortly we will be moving our entire 24/7 operation from this existing site to a completely new site and closing this existing site down.
The New site will have a min 40 meg leased line possibly 100meg, however it is all governed by the 2 meg existing leased line at the existing site.

We currently have 3 ESXi Hosts vers 5.1 with a variety of virtuals on each.
MY QUESTION IS
Some of the virtual guests cannot be simply shut down for a long period of time ie a weekend and the hosts transported over to the new site and booted back up as they are connected to 24/7 production machinery.
Is there a way that Veeam could be set to replicate these virtual guests over to the new sight bearing in mind some of them have over 1tb of data on them.

We can afford to shut these virtual guests down for approx. 30 mins possibly 1 hour but any longer and we are pushing our luck as production machinery is then down costing money.
We are trying to come up with the simplest and most effective way of having as little down time as possible in moving data from one site to another.
All the data is stored using HP Lefthand Sans.
If anyone has any other way of doing this either via Veeam or Vmware we are open to ideas. I was thinking of using a Veeam Replica backup over at the new site by
• creating the initial Veeam Replica at the existing site on the new ESXi Host/guest
• Taking that said host/guest and its storage over to the new sight ready to bring it up as the new master when we are ready.

If this is the best way great, if not we would really appreciate any ideas input.

Thanks

[foggy]

Roy, you're thinking right, replication is what you need here to guarantee minimum VMs downtime during the migration. Please look through this thread for more considerations. Btw, seeding the replica on such a slow line is also a good decision.

[royp]

Nice one Alexander,
Forgot to ask down about seeding glad you mentioned that as that was my next question, just noticed this thread after posting my question. I just got down to my post after reading all the above.
Seems straight forward enough, he says, famous last words eh.
Thanks for the quick response as usually Alexander

[foggy]

You're welcome. Feel free to ask for any additional clarification you need.

[emachabert]

Did this multiple times with success.
Largest migration was from Lyon to Paris over a leased L2 link @1gb. Moved 30TB of VM smoothly over a 3 month timeframe.
End users never noticed anything since cutover were done at night. The L2 link was also carrying 300 vlans.
Veeam is the best migration tool for vsphere to vsphere migrations.

[royp]

Thanks eric and Alexander for the input.

We can move most things with out worrying too much about down time within reason of course, however its the Centos servers which connect upto our Centos run machinery which runs 24/7 and takes quite a lot of timing to shut this machinery down correctly. Our file server / exchange / dcs / sql etc etc is not an issue hopefully should be relatively straight forward, he says touching wood.

If we decided to change the source servers IP range from say 192.9.200.* to 192.168.0.* for arguaments sake in the NEW DESTINATION when we bring it online, I notice the option for seperate virtual networks and differant IP addressing on the rep job.
HOWEVER
If for speed of the initial seed job we ran it on the existing site on the same 192.9.200.* then took it over to the new destination 192.168.0.* site, would I then be able to change the IP addressing options within the rep job to take into account that its then on a differant subnet. (just about to read up on this but always best to ask those that know it in side out)
This isnt a 100% definate, Im just trying to get my head around our options before hand, I would also create a few mickey mouse test rep jobs first to try the options out and see how things do and dont work, but just wanted to know what will and wont work with veeam.
Hopefully that makes sense.

[foggy]

HOWEVER
If for speed of the initial seed job we ran it on the existing site on the same 192.9.200.* then took it over to the new destination 192.168.0.* site, would I then be able to change the IP addressing options within the rep job to take into account that its then on a differant subnet.

Yes, the next restore point (the one created after mapping replication job to the replicas transferred offsite and specifying re-IP settings) will reflect the address changes.

[blodsbror]

Hi,

We have approximately 150 Hyper-v VM's that we need to migrate from one production environment to another (2012 R2 to 2012 R2 and different storage). I'm just wondering how Veeam could help us in this scenario ?

Is it possible to use veeam replication functionality to achieve this ? We could of course use SCVMM, however as Veeam is independent of the hypervisor and probably requires much less configuration (the two platforms are joined to two different domains), I'm just wondering if there could be any advantage in using it over SCVMM.

I've heard Veeam ZIP mentioned a lot, but I guess that's just single VM's at a time. Any tips would be appreciated!

thanks!