Replication

[ashman70]

I have some questions about replication. We are running Vsphere Essentials and have two hosts, I would like to put the second host at a second office and use it for replication. The second office is connected to the first (there is a third office too) via a 1.5MBps MPLS network. I would do the initial replication with both hosts on the same network, then transport the second host to the second office. Even though both offices are on different networks, they can still ping eachother and access network shares etc. Will this work for replication? Also, in the event of a disaster at the main office affecting the primary Esxi host, would I be able to bring up some or all of the replica's and have them function, even though they are on a different network? How does it work if I want to do this?

Thanks

[Vitaliy S.]

Yes, if secondary site is accessible from the main site you should be good. Make sure you have two proxy servers deployed on each site to make traffic transmission more efficient.

You will be able to bring your replicas up via vSphere Client or Veeam backup console (assuming it is installed on the DR site). BTW, since these sites have different networks, you can use RE-IP feature to make the failover process more flawless.

Let me know if you have any further questions. Thanks!

[tsightler]

While the IP networking is sufficient, the big concern here is the speed of your link. A 1.5Mbps link can only transfer 15GB of data total for the entire day, and that assumed you can use the entire link for the entire day, which is unlikely. How much data do you have and do you have any idea of your change rate? Those are the critical factors in determining the practicality of using replication for your case.

[ashman70]

I actually have quite a few questions.

One of the machines that will be replicated is an Exchange server, another is an application server running an MSSQL server and the other two are a terminal server. I am planning on getting the 1.5MBps link upgraded to 3.0 hopefully that will suffice, the link is really only used for terminal server connections so its utilization is not high.

How does constant replication work and would that be recommended for these servers?

Will I require a second copy of Veeam Backup and Replication to achieve what I am trying to do?

Can I use workstations as proxies or do they need to be servers, do they require fast disks or a lot of RAM?

If I have to bring up the Exchange server at the DR site, I am assuming it would not be able to send or receive external email since its gateway is at the primary site?

[Vitaliy S.]

I am planning on getting the 1.5MBps link upgraded to 3.0 hopefully that will suffice, the link is really only used for terminal server connections so its utilization is not high.

Not sure that even this bandwidth will allow you to replicate these VMs very often. How are you going to schedule these jobs?

How does constant replication work and would that be recommended for these servers?

Over this link you will always have an open snapshot if you're going to replicate these VMs. Try running replication jobs locally and see how many changes you will have with the desired job schedule.

Will I require a second copy of Veeam Backup and Replication to achieve what I am trying to do?

It depends, if you want to use re-ip and failback functions then you will need to protect your backup server as well (from losing the entire PROD site), and the best way to do that is to have a dedicated VBR server on the remote site.

Can I use workstations as proxies or do they need to be servers, do they require fast disks or a lot of RAM?

Please see our system requirements for proxy servers in the release notes document, but generally speaking proxy server can be any spare server you have, even the workstation would be fine.

If I have to bring up the Exchange server at the DR site, I am assuming it would not be able to send or receive external email since its gateway is at the primary site?

Yes, you have to provide protection for the gateway or have an alternative configuration on the remote site that would allow you to send/receive emails.

[ashman70]

What kind of bandwidth would you recommend then to replicate these VM's?

What kind of schedule would you recommend then, the idea is for disaster recovery failover.

Is it a bad thing to have a snapshot open all the time if you are doing constant replication?

[veremin]

What kind of bandwidth would you recommend then to replicate these VM's? What kind of schedule would you recommend then, the idea is for disaster recovery failover.

It depends on your needs and desires. As mentioned, you should first run replication jobs locally, understand how often you need to use them, and see what amount of data is generated between desired intervals. Based on the gotten information, you will be able to estimate bandwidth requirements.

Thanks.

[ashman70]

How does one ascertain this:

"understand how often you need to use them"

[foggy]

I guess Vladimir is talking about RPO value for replication jobs here, i.e. how far back you would need to be able to restore in case of DR (the acceptable amount of lost data). That defines the frequency of the job runs and hence, the amount of changes occurred inside the VM between those runs.

[veremin]

Yes, this is exactly what I've meant. The corresponding value is dictated solely by your business requirements, and no one apart from you and your boss can choose this value for you. Thanks.