Replicate to and from DataCenter

[lukejf]

Hey Guys
We have a customer that is looking at moving some services to a data center. They are also thinking about keeping some VM's onsite such as file servers etc. They currently run veeam and we are looking at setting up replication. Most of the replication will be from the datacenter to on premise as a DR site, however we will need to replicate some on premise VM's to the data center. I'm gathering two way replication is a pretty standard setup and can be easily achieved using veeam. We are recommending a dedicated 1Gbps network for the replication traffic, They have around 180GB per day in changes. We would also like to keep full backups of the datacenter vm's on premise that can be written to tape as a fail safe. Just checking this sounds like a sound solution and if any of you guys run anything similar.

[foggy]

Hi Luke, replication is a right approach in case of datacenter migration scenarios, consider using planned failover to provide minimum VMs downtime. Your other considerations also look good to me.

[skrause]

If the link between the data center is 1Gbps, if you have VMs that don't need the quicker recovery/failover that replication offers you might want to look at just targeting your backups to storage in the "on premise" location. Backups have much less limitations in terms of retention and they also take up less space on disk than replicas.

We have a change rate somewhat close to that in one of our datacenters that gets backed up in a similar fashion over a 1Gbps link nightly and don't have a whole lot of trouble with it.

I would also recommend, if you are running replication jobs for DR purposes, that you have those replication jobs managed by a Veeam server in the target location. Otherwise recovery becomes a lot more complicated. Veeam can share the proxy/repository components between multiple management servers so it would not require much more than setting up a second Veeam instance on a VM in the "other" location. If both of your Veeam servers are VMs you can also have them back each other up as an added bonus.

[lukejf]

Awesome info guys i really appreciate it. We will be do some testing in the next few weeks. I will let you know how it all goes.

[lukejf]

Guys another quick question, I'm scoping up the disks for my replication host and just wondering how you calculated the restore point size. We are currently using reverse incremental once a day. Is it simply the size of the vrb (changes). Are they stored on the target esxi server compress as in the VRB original file or would the restore point be larger than the original vrb.
Thanks
Hopefully this makes sense. I just want to ensure we have ample storage so dont run the risks of the replication jobs filling the hosts.

[skrause]

Replica VMs will take up the same amount of disk space as the original VMs + whatever the snapshots for restore points take up. Replication doesn't take advantage of deduplication or compression as the snapshots are regular VMware snapshots.

The snapshot disk space used will vary depending on how many restore points you are wanting and what your change rate is.

[lukejf]

thanks, do you have any tips for calculating change rate. can this info be found in the existing veeam backup jobs.

[jmmarton]

Veeam ONE has two reports that could help: VM Change Rate Estimation and VM Change Rate History.

Joe

[lukejf]

If the link between the data center is 1Gbps, if you have VMs that don't need the quicker recovery/failover that replication offers you might want to look at just targeting your backups to storage in the "on premise" location. Backups have much less limitations in terms of retention and they also take up less space on disk than replicas.

We have a change rate somewhat close to that in one of our datacenters that gets backed up in a similar fashion over a 1Gbps link nightly and don't have a whole lot of trouble with it.

I would also recommend, if you are running replication jobs for DR purposes, that you have those replication jobs managed by a Veeam server in the target location. Otherwise recovery becomes a lot more complicated. Veeam can share the proxy/repository components between multiple management servers so it would not require much more than setting up a second Veeam instance on a VM in the "other" location. If both of your Veeam servers are VMs you can also have them back each other up as an added bonus.

This has been very helpful thanks, I'm testing the setup as we speak in a lab and will let you know how we go. In regards to setting up a second full B+R server in the DR site for replication, What will happen if the replication and local on premise backup Job runs at the same time. Because they are managed on separate servers, will this cause an issue? Also is there a specific way to join the servers besides just connecting to the proxy service at each site on each server

[veremin]

Because they are managed on separate servers, will this cause an issue?

Job started second will fail.

Also is there a specific way to join the servers besides just connecting to the proxy service at each site on each server

All you need is to add a production host VMs to replicate reside on to a second backup server and assign proxy roles to two servers (one per site). That's it.

Thanks.

[lukejf]

thanks, appreciate it

[veremin]

You're welcome. If any additional help is needed, feel free to ask. Thanks.