I am looking into Veeam Backup & Replication and using the trial to replace Backup Exec for good. The reason I have opted for Veeam is the ability to 'connect to the cloud' as we are starting to migrate some of our servers into AWS.
Ultimately we want to be able to replicate from VMware vSphere directly into EC2 instances (real time!) but I believe it is not possible??? Please someone correct me if i'm wrong.
Instead I have been testing the following:
1. Backup a VM from our VMware vSphere environment (located on site) to our veeam server (also located on site)
2. Transfer this compressed backup to a windows server we have on AWS (over a VPN connection so cloud edition is not used at all)
3. I then extract using Veeam's extraction tool to uncompress
4. I then convert the extracted files to OVF format using VMware's tool
5. I then upload to our AWS VPC using Amazon's EC2 command line tools
This has resulted in our VMware VM backup running on AWS in an EC2 instance. This is good but the process is extremely long winded and i was wondering how it could be streamlined?
I want to be able to 'replicate' to AWS from our vSphere environment in as close to real time as possible but will settle for the next best thing if this cannot be done. I just don't quite know what that is at the moment. Can someone please help or point me in the right direction?
A mix of Windows / Linux machines, web servers, databases, ftp servers and a few specific applications (developed in house and 3rd party). Possibly even desktops at some point. I have AD replication working in the cloud currently. We want to ideally be able to replicate all servers that exist on site (primary) to AWS and that exist on AWS (primary) to on site for DR purposes.
Yes, I understand your desired DR design and it makes sense. I am just naturally interested in this other part of story
How do you determine which servers and workloads to run in Amazon, and which on premise? Is there a rule of thumb?
We are redesigning our infrastructure at the same time so are only moving servers or clusters of servers that have no to little communication requirements with anything that is still left on site because of the bandwidth limitation. Any servers we can split up/combine we do so, the servers that are left are the ones that will require the 'most work' and will probably not be moved for a year or two.
Louis,
regarind your opening post, there is little that can be done to shrink the process because Veeam supports VMware and Hyper-V, while AWS uses xen as the hypervisor, so at some point of the process, no matter how automated it is, there will always be a conversion between VM formats.
Is there a specific reason to choose AWS? Because otherwise you could look up to other cloud providers using VMware as the underlying hypervisor, in this way the conversion process could be avoided, so a large part of the time taken by the overall process could be removed.
Luca Dell'Oca Principal EMEA Cloud Architect @ Veeam Software
You might want to watch this video around automating the process. It doesn't really change the process at all, but it might help you to automate the conversion and certainly you can run tasks in parallel which could help.
