Environment Setup Advice Needed

[Andrew800]

Setup:

a - We have a Blade chassis currently with 3 ESXi hosts running 20+ VMs in an HA configuration (vCenter Server) using an iSCSI shared storage system. This equipment is set up at a remote datacenter managed from our headquarter office, therefore, 24-hr physical access is limited.

b - We also have another that will be similar to <a> and perhaps 2 other single-host ESXi environments that do not utilize vCenter Server.

At this time, our primary concern is getting the blade system squared away, but obviously the other systems are not far behind. We'd like to be able to run daily backups, and weekly full backups. We'd also like to have a disaster recovery contingency plan for offsite recovery (in case a missile hits the place). At the very least, even if we do not have matching equipment at another facility, we'd like to have the data available to be able to reconstruct the environment as it was prior to such a disaster, on top of having a local backup ready to go ASAP in case of only hardware failure or other local losses.

What configuration(s) or procedures can you suggest?

[*We do not yet have the product licensed yet - however, I just installed & briefly tested the 6.5 trial yesterday, while of course, today, v7 is released. I will try to upgrade to 7 (or uninstall 6.5 then install 7) as 6.5 is pointless now since we don't yet own either. I just downloaded the v7 trial of the Management Suite with Veeam One.]

[foggy]

At the very least, even if we do not have matching equipment at another facility, we'd like to have the data available to be able to reconstruct the environment as it was prior to such a disaster, on top of having a local backup ready to go ASAP in case of only hardware failure or other local losses.

There's no requirement to have similar equipment for successful utilization of Veeam B&R replication jobs. So setting up replication of your VMs to remote site will allow you to stand possible loss of the primary site and be able to fail over to replicas quickly (just make sure you have a proxy server deployed in the target location for optimal replication performance). Backup jobs to local repository will provide the ability for operational restores right in your production site.

[veremin]

We'd also like to have a disaster recovery contingency plan for offsite recovery (in case a missile hits the place).

Additionally, in case of replication, it might be worth deploying additional instance of VB&R in DR site and setting it responsible for replication jobs only. Having VB&R installed in DR location should guarantee that if any disaster happens, special operations like Failover or Failback can be started from there smoothly.

Thanks.

[Andrew800]

So are there clear-cut instructions on how to set this up? Say, for example, we have a large networked storage system at our headquarters (remote)... could/do we replicate the backup job (once completed locally or simultaneously) over a VPN to this storage? Must it hit a remote server running another instance of VB&R first (at the heaquarters)? If that's the case, how does licensing the software play into this?

[Basically, we want to backup up our primary site's VMs/data locally, then copy the backup data to a large hard drive or some other storage system at our remote headquarters, and have the ability to use that data (local or remote backup data) to either restore back into the functional environment (repair for data corruption, accidental deletes, failed disks, etc.), or rebuild the entire environment from the most recent backup point once we have the hardware back online (after a site disaster)]

[veremin]

This is the exact use case of new Backup Copy Job. It’s a new type of the job that will run continiously and handle the process of copying backup data off-site/DR location for you. All you need to do is add to DR location as a repository server to a backup console, create a backup copy job and specify required source (corresponding backup jobs) and target (DR) for it.

That’s, actually, all.

Thanks.

[Andrew800]

Thanks. I'll attempt to set this up but I'd also like to add one other point and a question: The VB&R software is installed and operates on a VM within this environment. I created a Windows 2k8R2 VM specifically to run as our backup server. How will this play into backups/recovery? Is it capable of backing itself up?

Also, as far as local storage space is concerned, what are the limits you can allocate for LUNs, VMs and repositories? Can I create a 2TB LUN, add a hard drive to this VM (backup server where VB&R is installed) and allocate that entire amount of space to it? What is the best way to determine how much space will be needed? [We have several VMs with 40 - 220GB drives allocated but are not near capacity thus far] What are best practices for this?

[foggy]

The VB&R software is installed and operates on a VM within this environment. I created a Windows 2k8R2 VM specifically to run as our backup server. How will this play into backups/recovery? Is it capable of backing itself up?

There is a huge existing topic regarding backup and replication of the Veeam B&R VM, please review it for answers.

Also, as far as local storage space is concerned, what are the limits you can allocate for LUNs, VMs and repositories? Can I create a 2TB LUN, add a hard drive to this VM (backup server where VB&R is installed) and allocate that entire amount of space to it? What is the best way to determine how much space will be needed? [We have several VMs with 40 - 220GB drives allocated but are not near capacity thus far] What are best practices for this?

Space required to store your backups is hardly predictable as depends on too many factors (number of VMs, amount of changes VMs generate between backup cycles, required retention policy, etc.). So you better test backups during a week, for example, and extrapolate the results to your planned retention.

Another point that should be made is that storing backup on VMFS is not considered to be a best practice.

[veremin]

What are best practices for this?

Additionally, if you’re concerned with storage space saving, I’d recommend using reversed incremental mode, since it consumes less space compared to forward incremental method. Thanks.